Q2E: Query-to-Event Decomposition for Zero-Shot Multilingual
Collection
Datasets used in the paper: Q2E: Query-to-Event Decomposition for Zero-Shot Multilingual Text-to-Video Retrieval • 37 items • Updated
query stringclasses 259 values | prequel listlengths 0 30 | sequel listlengths 0 30 | during listlengths 0 30 | metadata dict | num_of_frames int64 16 16 | frame_captions listlengths 16 16 | video_id stringlengths 6 45 | frame2video_caption stringlengths 520 3.4k | asr dict |
|---|---|---|---|---|---|---|---|---|---|
The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea. | [
"People running out of buildings during the 2016 Gyeongju, South Korea earthquake on September 12",
"Emergency responders searching through rubble after the 2016 Gyeongju, South Korea earthquake on September 12, 2016",
"Cars driving away from the affected area in Gyeongju, South Korea after the earthquake on 12... | [
"Buildings collapsing or showing significant structural damage during the 12 September 2016 earthquake in Gyeongju, South Korea",
"People evacuating their homes and gathering in safe areas in Gyeongju, South Korea after the earthquake on September 12, 2016",
"Emergency responders searching for survivors in rubb... | [
"People running out of buildings during the earthquake in Gyeongju, South Korea on September 12, 2016",
"Rescue teams searching through rubble after the 12 September 2016 earthquake in Gyeongju, South Korea",
"Buildings with cracked walls and fallen debris in Gyeongju, South Korea after the 12 September 2016 ea... | {
"article_url": "https://ko.wikipedia.org/wiki/2016%EB%85%84_%EA%B2%BD%EC%A3%BC_%EC%A7%80%EC%A7%84",
"during_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to decompose the ongoing event in multiple simple events. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"during_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the consequences and aftermath of an earthquake, as these are the most likely to be visualized in a video. Events such as damage to buildings, rescue operations, and people reacting to the quake are plausible and commonly seen in earthquake-related videos.\n\nEVENTS:\n1. People running out of buildings\n2. Rescue teams searching through rubble\n3. Buildings with cracked walls and fallen debris\n4. Emergency vehicles rushing to the scene\n5. People gathering in open spaces for safety",
"en_article_excerpt": "The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.",
"en_article_url": "https://en.wikipedia.org/wiki/2016_Gyeongju_earthquake",
"event_category": "disasters",
"event_name": "gyeongju_earthquake",
"language": "korean",
"prequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what previous events could lead to this given event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"prequel_raw_generation": "EXPLANATION: It makes sense to extract events that are directly related to the aftermath or impact of an earthquake, as these would be the most visually apparent and relevant to a video about the 2016 Gyeongju earthquake. Events such as destruction, evacuation, and rescue efforts are likely to be seen in a video associated with this query.\n\nEVENTS:\n1. People running out of buildings\n2. Emergency responders searching through rubble\n3. Cars driving away from the affected area\n4. Buildings with cracked walls and broken windows\n5. Rescue teams carrying injured people on stretchers",
"sequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what events can be the outcome of this ongoing event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"sequel_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the aftermath and impact of an earthquake, such as damage to infrastructure, rescue efforts, and evacuation procedures. These events are likely to be visually represented in a video associated with the query, providing a concrete and tangible depiction of the earthquake's effects.\n\nEVENTS:\n1. Buildings collapsing or showing significant structural damage\n2. People evacuating their homes and gathering in safe areas\n3. Emergency responders searching for survivors in rubble\n4. Roads blocked by debris and damaged vehicles\n5. Residents assessing damage to their homes and belongings",
"video_URL": "http://youtube.com/watch?v=t8lY-rjy7-Y",
"video_description": "9월 12일, 경주에서는 역대 최대 규모인 5.8 지진이 일어났다. 지진으로 서울과 수도권까지 전국이 흔들렸고, 이로 인해 전화와 카카오톡도 모두 먹통이었다. 정부는 지진 발생 9분이 지나서야 긴급재난문자를 발송했고 이마저도 서울과 수도권 지역 국민들은 문자를 받지 못했다. #경주 #지진 #재난문자",
"video_id": "t8lY-rjy7-Y"
} | 16 | [
"The image depicts a group of people, primarily women, rushing through an indoor setting. They appear to be in a state of panic or urgency, moving quickly down a corridor and up a staircase. The text \"어제 저녁 7시 44분\" (Yesterday at 7:44 PM) suggests that the scene is from a specific time in the past. The individuals... | t8lY-rjy7-Y | The video depicts a sequence of scenes related to disaster preparedness and response, primarily focusing on an earthquake scenario. It begins with a chaotic scene of people, mostly women, rushing through an indoor setting, indicating a state of panic or urgency, possibly in response to an earthquake warning or event. This is followed by a classroom scene where students are hastily gathering their belongings and evacuating, suggesting an emergency drill or a real earthquake event.
The video then transitions to a train station platform at night, which appears calm and empty, contrasting with the previous chaotic scenes. This is followed by scenes of well-organized storage areas, suggesting preparedness and orderliness in anticipation of future events.
However, the video also shows scenes of disruption and chaos, such as a street with broken tiles and debris, indicating the aftermath of an earthquake. A nighttime scene of a bus navigating through a debris-filled road underscores the challenges of moving through affected areas.
The second half of the video shifts focus to official responses and preparations. It includes scenes of a person attending to someone on the floor after a quake, a government webpage announcing support for affected areas, and official briefings or press conferences where individuals are discussing disaster management and the importance of preparedness.
The video concludes with scenes of students in a school setting, possibly participating in an emergency drill, and a formal meeting between high-ranking officials, including the President and Prime Minister, discussing governance and policy issues.
Throughout the video, the theme of preparedness for future earthquakes and other disasters is emphasized, along with the importance of governmental and community responses to such events. The original ASR transcript, "I hope that we will be prepared for a bigger earthquake in the future," encapsulates the overarching message of the video, which is to stress the need for thorough preparation and effective response mechanisms to mitigate the impact of natural disasters. | {
"original": " I hope that the earthquake will be prepared for a bigger one in the future.",
"refined": "I hope that we will be better prepared for a bigger earthquake in the future.",
"summarized": "The speaker expresses a hope that preparations will be in place for a potentially larger earthquake in the future, and highlights the importance of predicting and preparing for great disasters, possibly referring to a specific event expected in 5 years.",
"translated": "I hope that we will be prepared for a bigger earthquake in the future.",
"translated_llm": "I hope that the earthquake will be prepared for a bigger one in the future.",
"translated_whisper": " The The The The The The The The The The The The The The The The The The The The The The The The The The The The The ... ... ... ... ... ... ... ... ... ... ... ... ... ... Even the most difficult thing is A huge disaster The disaster is coming And now we need more Great disasters to be predicted A new emergency in 5 years That day Now This place will be"
} |
The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea. | [
"People running out of buildings during the 2016 Gyeongju, South Korea earthquake on September 12",
"Emergency responders searching through rubble after the 2016 Gyeongju, South Korea earthquake on September 12, 2016",
"Cars driving away from the affected area in Gyeongju, South Korea after the earthquake on 12... | [
"Buildings collapsing or showing significant structural damage during the 12 September 2016 earthquake in Gyeongju, South Korea",
"People evacuating their homes and gathering in safe areas in Gyeongju, South Korea after the earthquake on September 12, 2016",
"Emergency responders searching for survivors in rubb... | [
"People running out of buildings during the earthquake in Gyeongju, South Korea on September 12, 2016",
"Rescue teams searching through rubble after the 12 September 2016 earthquake in Gyeongju, South Korea",
"Buildings with cracked walls and fallen debris in Gyeongju, South Korea after the 12 September 2016 ea... | {
"article_url": "https://ko.wikipedia.org/wiki/2016%EB%85%84_%EA%B2%BD%EC%A3%BC_%EC%A7%80%EC%A7%84",
"during_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to decompose the ongoing event in multiple simple events. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"during_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the consequences and aftermath of an earthquake, as these are the most likely to be visualized in a video. Events such as damage to buildings, rescue operations, and people reacting to the quake are plausible and commonly seen in earthquake-related videos.\n\nEVENTS:\n1. People running out of buildings\n2. Rescue teams searching through rubble\n3. Buildings with cracked walls and fallen debris\n4. Emergency vehicles rushing to the scene\n5. People gathering in open spaces for safety",
"en_article_excerpt": "The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.",
"en_article_url": "https://en.wikipedia.org/wiki/2016_Gyeongju_earthquake",
"event_category": "disasters",
"event_name": "gyeongju_earthquake",
"language": "korean",
"prequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what previous events could lead to this given event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"prequel_raw_generation": "EXPLANATION: It makes sense to extract events that are directly related to the aftermath or impact of an earthquake, as these would be the most visually apparent and relevant to a video about the 2016 Gyeongju earthquake. Events such as destruction, evacuation, and rescue efforts are likely to be seen in a video associated with this query.\n\nEVENTS:\n1. People running out of buildings\n2. Emergency responders searching through rubble\n3. Cars driving away from the affected area\n4. Buildings with cracked walls and broken windows\n5. Rescue teams carrying injured people on stretchers",
"sequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what events can be the outcome of this ongoing event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"sequel_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the aftermath and impact of an earthquake, such as damage to infrastructure, rescue efforts, and evacuation procedures. These events are likely to be visually represented in a video associated with the query, providing a concrete and tangible depiction of the earthquake's effects.\n\nEVENTS:\n1. Buildings collapsing or showing significant structural damage\n2. People evacuating their homes and gathering in safe areas\n3. Emergency responders searching for survivors in rubble\n4. Roads blocked by debris and damaged vehicles\n5. Residents assessing damage to their homes and belongings",
"video_URL": "http://youtube.com/watch?v=CXe6_CDrdB8",
"video_description": "찐♡기상청사람들 예고 없이 찾아오는 지진! 한 번 발생하면 큰 피해를 주는 지진이기에 감시하는 일은 한시도 소홀히 할 수 없습니다. 그 최전방에서... 1년 365일 24시간 지진화산 감시 및 분석 업무에 최선을 다해 신속·정확하게 알리고자 노력하는 사람이 있습니다. 지진화산 감시 및 분석 업무 9년 차... 반복되는 주간 근무와 야간 근무가 힘들긴 하지만, 누구보다 묵묵하게 자리를 지키고 있는 진득한 뚝배기 같은 지진업무의 대들보!! 지진화산 감시 업무를 담당하는 기상청 사람을 만나고 왔습니다.(^_^)v 우리나라에서 가장 강력했던 규모 5.8의 경주 지진 때 지진감시 및 분석을 수행했던 기상청 사람은 누구? 지진감시 업무를 들어올 때마다 긴장되는 기상청 사람은 누구? 국민의 안전을 누구보다 걱정하는 기상청 사람은 누구? 지진 올바로 알고 지진 발생 시 빠르게 대처하길 바라는 기상청 사람은 누구? 최근 지진 안전을 홍보해 주시는 지진희 배우님을 그 누구보다 고마워하는 기상청 사람은 누구? 만나볼까요 :) #기상청사람들 #지진화산 #행동하는파수꾼",
"video_id": "CXe6_CDrdB8"
} | 16 | [
"The current frame shows a person standing in front of a control room display screen filled with data and graphs. The individual appears to be speaking about earthquake monitoring and reporting, mentioning their experience with analyzing a significant earthquake in the economic sector. The text on the screen indica... | CXe6_CDrdB8 | The video depicts the work of an earthquake analyst in a control room setting, focusing on the monitoring, analysis, and reporting of seismic events. The analyst, dressed in a blue shirt, is shown working at a desk with multiple computer monitors displaying data and graphs related to earthquake activity. The video highlights the importance of swift and accurate action in earthquake reporting, as the analyst works to analyze and report on earthquakes, including a significant event with a magnitude of 5.8 that occurred in the economic sector. The analyst's work involves collaboration with others, as seen in frames where they discuss earthquake data with colleagues, and they also address the audience directly, emphasizing the need for rapid and precise seismic analysis. The video emphasizes the unpredictability of earthquakes and the resulting anxiety and tension faced by those responsible for monitoring and reporting them. Throughout the video, the analyst's role is portrayed as critical in providing timely and accurate information about seismic events, and they express gratitude for the opportunity to work in this field. The overall context is one of professionalism, urgency, and dedication to ensuring public safety through effective earthquake monitoring and response. | {
"original": " 국내 뿐만 아니라 국외에서 발생하는 지진을 분석을 해서 일정 규모 이상이 되면 통보를 하고 있고요. 더불어서 화산 정보를 제공을 하고 있습니다. 경제에서 발생한 지진인데요. 아찔한데요. 당시만 해도 규모가 큰 지진은 별로 없었거든요. 분석하고 있는 분은 엄청 떨리고 식은땀은 나고 일단 불안감이 먼저 앞섰습니다. 1.1의 지진을 통보를 했는데 40분, 50분 뒤에 앞서 발표한 지진보다 훨씬 큰 게 감이 오더라고요. 최대한 더 빨리 해야 되겠다 분석을 시작했는데 규모 5.8이 나왔어요. 충분히 만족을 하고 있습니다. 지진이 어느 시각에 발생할지 모르기 때문에 항상 불안하고 긴장감을 가지고 근무하고 있지만 지진을 맡고 있는 담당자의 숙명이라고 생각합니다. 지진이 배우님, 지진 업무를 담당하고 있는 사람으로서 고마움을 느끼고 있습니다. 화이팅입니다. 지진, 올바로 알아야 빠르게 대처할 수 있습니다. 지진은 신속정학입니다. 제가 그 신속정학의 중심에 서겠습니다",
"refined": "In Korea and abroad, we analyze earthquakes and report those that exceed a certain magnitude. Additionally, we provide volcanic information. I'm referring to the earthquake that occurred in the economy. It was quite alarming. At the time, there weren't many large earthquakes, so the analyst was extremely nervous and anxious. After reporting a 1.1 magnitude earthquake, we felt that a much larger one was coming 40 to 50 minutes later. We began analyzing as quickly as possible and determined it was a 5.8 magnitude earthquake. I'm satisfied with the results. Since earthquakes can occur at any time, I'm always anxious and on edge, but I consider it an occupational hazard. As someone in charge of earthquake duty, I feel grateful. I'll do my best to provide accurate information, as understanding earthquakes correctly is crucial for prompt action.",
"summarized": "The speaker is an earthquake analyst who monitors and reports earthquakes in Korea and other countries. They recall a particular earthquake in Gyeongju that initially seemed small, but later turned out to be much larger, with a magnitude of 5.8. The analyst expresses satisfaction with their work and acknowledges the anxiety and tension that comes with the job, but feels a sense of responsibility and gratitude for being able to contribute to public safety through their work. The speaker emphasizes the importance of accurate and rapid information about earthquakes to facilitate quick response and preparedness.",
"translated": "We analyze earthquakes not only in Korea but also in other countries, and if they exceed a certain magnitude, we report them. Additionally, we provide volcanic information. This is about an earthquake that occurred in the economic sector, which was quite shocking. At the time, there weren't many large earthquakes, so the person analyzing it was extremely nervous, sweating, and felt a great deal of anxiety. We reported an earthquake with a magnitude of 1.1, but 40 to 50 minutes later, it became clear that it was going to be much larger than initially announced. We began analyzing as quickly as possible and determined that the magnitude was 5.8. I'm satisfied with the result. Since we can't predict when an earthquake will occur, we always work with a sense of anxiety and tension, but I believe this is the nature of being in charge of earthquake duty. As someone responsible for earthquake work, I feel grateful. Earthquakes require swift and accurate action, so it's essential to understand them correctly. I will be at the forefront of this rapid response.",
"translated_llm": "The earthquake is a major economic event. It is a painful event. It was not a major earthquake at the time. The person who was analyzing it was very shaken and sweating and I was very anxious. I reported a 1.1 earthquake, but after 40 minutes, 50 minutes, I was told that it was going to be much bigger than the one I had announced earlier. I started the analysis as soon as possible and it was 5.8. I am satisfied.",
"translated_whisper": " What is the earthquake in Korea? We analyze earthquakes in Korea and other countries and if they are of a certain size, we report them. In addition, we provide damage information. This is an earthquake from Gyeongju. It's a bit dizzy. At that time, there weren't many large earthquakes. I was very nervous and cold when I was analyzing. I was worried at first. I reported an earthquake at 1.2. After 40 to 50 minutes, I felt that it was much bigger than the previous earthquake. I started analyzing and I got a 5.8 scale. Q. How do you feel about the earthquake? I'm satisfied with the results. I don't know what the earthquake will look like, so I'm always anxious and nervous. But I think it's the life of the person in charge of the earthquake. Q. How do you feel about the earthquake? As an earthquake actor, I feel grateful as a person in charge of the earthquake. I'm cheering for you. Q. What is the earthquake? You have to know the earthquake correctly to react quickly. The earthquake is a rapid earthquake. I will be the center of that rapid development."
} |
The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea. | [
"People running out of buildings during the 2016 Gyeongju, South Korea earthquake on September 12",
"Emergency responders searching through rubble after the 2016 Gyeongju, South Korea earthquake on September 12, 2016",
"Cars driving away from the affected area in Gyeongju, South Korea after the earthquake on 12... | [
"Buildings collapsing or showing significant structural damage during the 12 September 2016 earthquake in Gyeongju, South Korea",
"People evacuating their homes and gathering in safe areas in Gyeongju, South Korea after the earthquake on September 12, 2016",
"Emergency responders searching for survivors in rubb... | [
"People running out of buildings during the earthquake in Gyeongju, South Korea on September 12, 2016",
"Rescue teams searching through rubble after the 12 September 2016 earthquake in Gyeongju, South Korea",
"Buildings with cracked walls and fallen debris in Gyeongju, South Korea after the 12 September 2016 ea... | {
"article_url": "https://ko.wikipedia.org/wiki/2016%EB%85%84_%EA%B2%BD%EC%A3%BC_%EC%A7%80%EC%A7%84",
"during_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to decompose the ongoing event in multiple simple events. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"during_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the consequences and aftermath of an earthquake, as these are the most likely to be visualized in a video. Events such as damage to buildings, rescue operations, and people reacting to the quake are plausible and commonly seen in earthquake-related videos.\n\nEVENTS:\n1. People running out of buildings\n2. Rescue teams searching through rubble\n3. Buildings with cracked walls and fallen debris\n4. Emergency vehicles rushing to the scene\n5. People gathering in open spaces for safety",
"en_article_excerpt": "The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.",
"en_article_url": "https://en.wikipedia.org/wiki/2016_Gyeongju_earthquake",
"event_category": "disasters",
"event_name": "gyeongju_earthquake",
"language": "korean",
"prequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what previous events could lead to this given event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"prequel_raw_generation": "EXPLANATION: It makes sense to extract events that are directly related to the aftermath or impact of an earthquake, as these would be the most visually apparent and relevant to a video about the 2016 Gyeongju earthquake. Events such as destruction, evacuation, and rescue efforts are likely to be seen in a video associated with this query.\n\nEVENTS:\n1. People running out of buildings\n2. Emergency responders searching through rubble\n3. Cars driving away from the affected area\n4. Buildings with cracked walls and broken windows\n5. Rescue teams carrying injured people on stretchers",
"sequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what events can be the outcome of this ongoing event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"sequel_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the aftermath and impact of an earthquake, such as damage to infrastructure, rescue efforts, and evacuation procedures. These events are likely to be visually represented in a video associated with the query, providing a concrete and tangible depiction of the earthquake's effects.\n\nEVENTS:\n1. Buildings collapsing or showing significant structural damage\n2. People evacuating their homes and gathering in safe areas\n3. Emergency responders searching for survivors in rubble\n4. Roads blocked by debris and damaged vehicles\n5. Residents assessing damage to their homes and belongings",
"video_URL": "http://youtube.com/watch?v=uF3x63vtddc",
"video_description": "\"경주 지진, 지하수는 1~2일 전에 미리 알았다\" 지난 12일과 19일 경주에서 규모 5.8을 비롯한 대형 지진이 잇따라 발생하기 하루 이틀 전 주변 지하수 수위가 크게 상승해, 향후 지진예측에 활용할 수 있다는 주장이 나왔습니다. 정상용 부경대 지구환경과학과 교수는 진앙에서 가장 가까운 경북 경주시 산내면에 있는 지하수 관측소 자료를 분석한 결과, 이 같이 밝혔습니다. 자료를 보면 지난 11일 암반 지하수 수위는 하루 사이에 무려 49㎝ 상승했고, 17일에도 39㎝나 올라갔습니다. 정 교수는 지진에 앞서 암석에 강한 압력이 가해져 지하수 수위가 많이 올라간 것으로 보인다며, 규모 4.5 이상의 지진에만 영향을 받았다고 설명했습니다. 연합뉴스TV : 02-398-4409(제보) 4441(기사문의), 카톡/라인 jebo23 (끝)",
"video_id": "uF3x63vtddc"
} | 16 | [
"The current frame features two news anchors in a television news studio, presenting a report. The text at the bottom of the screen in Korean states, \"경주 지진, 지하수는 1~2일 전에 미리 알았다\" which translates to \"Gyeongju earthquake, groundwater knew 1-2 days in advance.\" The time displayed is 09:34. The news program appear... | uF3x63vtddc | The video discusses the recent earthquakes in Gyeongju, particularly focusing on the potential predictive power of rising groundwater levels before seismic events. It begins with news anchors presenting a report on the topic, highlighting research by Professor Jeong Sang-yong of Pukyong National University. The professor analyzed data from an underground water observation station near the epicenter, finding significant increases in water levels days before the earthquakes. The footage then transitions through various scenes, including nighttime shots of heavy rainfall, a dramatic event of water gushing out of a structure, and calm urban street scenes, all emphasizing the connection between groundwater levels and seismic activity. The narrative is reinforced by text overlays translating to "Gyeongju earthquake, the groundwater was known a day or two in advance." Later frames show indoor and outdoor settings, including a construction scene where workers are repairing a traditional building's roof, possibly in response to earthquake damage. Throughout, the video maintains a focus on the potential for groundwater level changes to predict earthquakes, referencing the recent Gyeongju earthquakes as a case study. The original ASR transcript provides context, explaining that groundwater levels rose significantly before the earthquakes, with Professor Jeong suggesting this could be used for prediction. The video concludes by highlighting the importance of this research and the ongoing efforts to understand and prepare for seismic events, including the preservation of historic structures. | {
"original": " 지난 11일과 19일 경주에서 규모 5.8을 비롯한 대형 지진이 잇따라 발생하기 하루 이틀 전, 주변 지하수의 수위가 크게 상승해 향후 지진 예측에 활용할 수 있다는 주장이 나왔습니다. 정상용 부경대 지구환경과학과 교수는 진앙에서 가장 가까운 경북 경주시 산내면에 있는 지하수 관측소 자료를 분석한 결과 이같이 밝혔습니다. 자료를 보면 지난 11일 안반 지하수 수위는 하루 사이에 무려 49cm가 상승했고, 17일에도 39cm나 올라갔습니다. 정 교수는 지진에 앞서 암석에 강한 압력이 가해져 지하수의 수위가 많이 올라간 것으로 보인다며 규모 4.5 이상의 지진에만 영향을 받았다고 설명했습니다.",
"refined": "On the 11th and 19th, large earthquakes, including a 5.8-magnitude quake, occurred in Gyeongju. A day or two before the earthquakes, the groundwater level in the surrounding area rose significantly, which could be useful for predicting future earthquakes. Professor Jung of the Department of Earth and Environmental Science at Pukyong National University analyzed data from a groundwater observation facility in Sanin-myeon, Gyeongju, closest to the epicenter. The data shows that the groundwater level in Anban rose by 49cm in one day on the 11th and by 39cm on the 17th. According to Professor Jung, strong pressure on the rock prior to the earthquake caused a significant rise in the groundwater level, which was only affected by earthquakes with a magnitude of 4.5 or higher.",
"summarized": "Research by Professor Jung of the Earth Environment Science department at Pukyong National University suggests that changes in underground water levels may be used to predict future earthquakes. Analyzing data from an underground water observation facility in Gyeongju, the professor found that the water level rose significantly, 49cm on the 11th and 39cm on the 17th, before recent large earthquakes in the area. According to Professor Jung, the increased water level is likely due to strong pressure on rocks before an earthquake, and this phenomenon is only observed with earthquakes of magnitude 4.5 or higher.",
"translated": "A day or two before the recent large earthquakes, including a 5.8-magnitude quake, struck Gyeongju on the 11th and 19th, the groundwater level in the surrounding area rose significantly, prompting claims that this phenomenon could be used to predict future earthquakes. Professor Jeong Sang-yong of the Department of Earth and Environmental Science at Pukyong National University analyzed data from an underground water observation station in Sanneom-myeon, Gyeongju, which is closest to the epicenter, and reported these findings. According to the data, the groundwater level in Anban rose by 49cm in just one day on the 11th and by 39cm on the 17th. Professor Jeong explained that the increase in groundwater level appeared to be caused by strong pressure on the rocks before the earthquake and that only earthquakes with a magnitude of 4.5 or higher had an impact.",
"translated_llm": "The data from the nearest underground water observatory in the mountains of North Korea, located in the heart of North Korea, showed that the underground water level rose by 49 cm in the past 11 days and rose by 39 cm on the 17th. The professor explained that the underground water level appeared to have risen a lot before the earthquake and that the earthquake was only affected by a strong pressure on the rocks.",
"translated_whisper": " On the 11th and 19th, a large earthquake of 5.8 in Gyeongju, two days before the earthquake, The number of underground water around the area has risen significantly, and it is estimated that it can be used to predict the future. Professor of Earth Environment Science at the National University of Gyeongju, Gyeongsang-do, The data of the underground water observation facility in the mountains in Gyeongju, which is the closest to the Jinhang, has been analyzed. According to the data, the underground water level in Anban was 49cm higher in a day on the 11th, and it rose to 39cm on the 17th. Professor Jung said that the strong pressure of the rock was applied to the earthquake, and the underground water level rose a lot, and that it was only affected by an earthquake of 4.5 or more."
} |
The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea. | [
"People running out of buildings during the 2016 Gyeongju, South Korea earthquake on September 12",
"Emergency responders searching through rubble after the 2016 Gyeongju, South Korea earthquake on September 12, 2016",
"Cars driving away from the affected area in Gyeongju, South Korea after the earthquake on 12... | [
"Buildings collapsing or showing significant structural damage during the 12 September 2016 earthquake in Gyeongju, South Korea",
"People evacuating their homes and gathering in safe areas in Gyeongju, South Korea after the earthquake on September 12, 2016",
"Emergency responders searching for survivors in rubb... | [
"People running out of buildings during the earthquake in Gyeongju, South Korea on September 12, 2016",
"Rescue teams searching through rubble after the 12 September 2016 earthquake in Gyeongju, South Korea",
"Buildings with cracked walls and fallen debris in Gyeongju, South Korea after the 12 September 2016 ea... | {
"article_url": "https://ko.wikipedia.org/wiki/2016%EB%85%84_%EA%B2%BD%EC%A3%BC_%EC%A7%80%EC%A7%84",
"during_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to decompose the ongoing event in multiple simple events. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"during_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the consequences and aftermath of an earthquake, as these are the most likely to be visualized in a video. Events such as damage to buildings, rescue operations, and people reacting to the quake are plausible and commonly seen in earthquake-related videos.\n\nEVENTS:\n1. People running out of buildings\n2. Rescue teams searching through rubble\n3. Buildings with cracked walls and fallen debris\n4. Emergency vehicles rushing to the scene\n5. People gathering in open spaces for safety",
"en_article_excerpt": "The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.",
"en_article_url": "https://en.wikipedia.org/wiki/2016_Gyeongju_earthquake",
"event_category": "disasters",
"event_name": "gyeongju_earthquake",
"language": "korean",
"prequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what previous events could lead to this given event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"prequel_raw_generation": "EXPLANATION: It makes sense to extract events that are directly related to the aftermath or impact of an earthquake, as these would be the most visually apparent and relevant to a video about the 2016 Gyeongju earthquake. Events such as destruction, evacuation, and rescue efforts are likely to be seen in a video associated with this query.\n\nEVENTS:\n1. People running out of buildings\n2. Emergency responders searching through rubble\n3. Cars driving away from the affected area\n4. Buildings with cracked walls and broken windows\n5. Rescue teams carrying injured people on stretchers",
"sequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what events can be the outcome of this ongoing event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"sequel_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the aftermath and impact of an earthquake, such as damage to infrastructure, rescue efforts, and evacuation procedures. These events are likely to be visually represented in a video associated with the query, providing a concrete and tangible depiction of the earthquake's effects.\n\nEVENTS:\n1. Buildings collapsing or showing significant structural damage\n2. People evacuating their homes and gathering in safe areas\n3. Emergency responders searching for survivors in rubble\n4. Roads blocked by debris and damaged vehicles\n5. Residents assessing damage to their homes and belongings",
"video_URL": "http://youtube.com/watch?v=nln3QFJggKs",
"video_description": "106억 → 58억…경주 지진 피해금액 반토막 잇단 강진으로 100억원이 넘는 것으로 추정됐던 경주시의 공식 피해규모가 58억원 수준에 그치는 것으로 집계됐습니다. 경주시가 지진 발생 뒤 전수조사를 통해 발표한 잠정 추산 피해액은 약 107억원이었지만 사유재산 피해는 대부분 공식 피해액에서 빠졌습니다. 규정상 기와가 떨어지거나 벽체에 금이 가고 담장이 파손된 것은 재산피해 산정되지 않습니다. 이에 따라 사유재산 중에는 지진으로 주택이 완전히 부서지거나 반쯤 허물어진 것만 공식 피해액에 들어갔습니다. 연합뉴스TV : 02-398-4409(제보) 4441(기사문의), 카톡/라인 jebo23 (끝)",
"video_id": "nln3QFJggKs"
} | 16 | [
"The image shows two news anchors in a television studio, with a map of Korea displayed on the screen. The male anchor is dressed in a dark suit and tie, while the female anchor wears a blue blazer over a pink top. The text on the screen indicates that the official damage estimate for Gyeongju following an earthqua... | nln3QFJggKs | The video depicts the aftermath of an earthquake in Gyeongju, South Korea, focusing on the damage assessment and recovery efforts. It begins with a news report in a television studio, where anchors discuss the revised official damage estimate, which has been halved from 10.6 billion won to 5.8 billion won. The majority of private property damages, such as cracks in walls and broken fences, were not included in the official tally. The video then transitions to aerial views of the city, showing densely populated areas and open green spaces, highlighting the scope of the affected region.
The footage continues with scenes of demolition, reconstruction, and repair efforts, including a large crane lifting a portion of a building and workers manually repairing a traditional tiled roof. These scenes underscore the scale of the response and the physical labor involved in rebuilding. The video also shows streets littered with debris, emphasizing the need for thorough cleanup and repair work.
Further scenes depict individuals assessing damage to buildings, including a traditional structure with a severely damaged wooden ceiling and a man inspecting a cracked wall. Close-up views of cracked pavements, roads, and walls highlight the extent of the infrastructure damage. The text "지진 공식 피해금액 반토막" (official earthquake damage halved) is consistently displayed throughout, reinforcing the reduction in damage estimates.
The latter part of the video focuses on the detailed assessment of damage to residential properties and cultural heritage sites, with a mention of an emergency inspection of cultural properties following the earthquake. The interior of a building shows significant structural damage, including cracks in the ceiling and possibly water damage. A wall with missing tiles exposing underlying materials further emphasizes the impact on infrastructure.
Overall, the video provides a comprehensive view of the earthquake's aftermath in Gyeongju, from the initial news report to the ongoing recovery and assessment efforts, highlighting the challenges in evaluating and addressing the damage to both private properties and cultural heritage sites. | {
"original": " 이딴 강진으로 100억 원이 넘는 것으로 추정됐던 경주시의 공식 피해 규모가 58억 원 수준에 그치는 것으로 집계됐습니다. 경주시가 지진 발생 뒤 전수조사를 통해 발표한 잠정추산 피해액은 약 107억 원이었지만 사유재산 피해는 대부분 공식 피해액에서 빠졌습니다. 규정상 기화가 떨어지거나 벽체에 금이 가고 담장이 파손된 것은 재산 피해로 산정되지 않습니다. 사유재산 중에는 지진으로 주택이 완전히 부서지거나 반쯤 허물어진 것만 공식 피해에게 들어갔습니다.",
"refined": "The official damage in Gyeongju, initially estimated to exceed 10 billion won, has been recalculated to around 5.8 billion won. Following a thorough investigation after the earthquake, the city reported a preliminary estimated damage of approximately 10.7 billion won, but most private property damage was excluded from the official total. According to regulations, damage such as cracks in walls or broken fences is not considered property damage. Only private properties that were completely destroyed or partially demolished due to the earthquake were included in the official damage report.",
"summarized": "Gyeongju city's official damage from the earthquake was estimated to be over 1 billion won, but the final calculation is around 580 million won. The initial estimate of 1.07 billion won was reduced as most private property damages, such as cracks in walls or broken fences, were not included due to regulations. Only severe cases of private property damage, such as completely or partially destroyed houses, were accounted for in the official damage total.",
"translated": "The official damage estimate for Gyeongju, initially expected to exceed 10 billion won, has been revised to approximately 5.8 billion won. Following a thorough survey after the earthquake, the city reported preliminary estimated damages of around 10.7 billion won. However, most private property damages were not included in the official tally. According to regulations, damages such as cracks in walls, broken fences, or decreased building foundations are not considered property damage. Only cases where homes were completely or partially destroyed due to the earthquake were counted in the official damage estimate.",
"translated_llm": "The official damage estimated by the city of 경주 to be over 10 billion won was estimated to be at 58 billion won. The estimated damage was about 107 billion won, but most of the damage was lost from the official damage.",
"translated_whisper": " The official damage of Gyeongju, which was estimated to be over 1 billion won, has reached the level of 5.8 billion won. The total damage of the dam was about 1.7 billion won, which was announced through a full-scale survey after the earthquake. But most of the damage to the property was removed from the official damage. According to the regulations, the damage to the dam is not determined as damage to the property. Accordingly, the property was damaged by an earthquake, or half of the property was destroyed."
} |
The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea. | [
"People running out of buildings during the 2016 Gyeongju, South Korea earthquake on September 12",
"Emergency responders searching through rubble after the 2016 Gyeongju, South Korea earthquake on September 12, 2016",
"Cars driving away from the affected area in Gyeongju, South Korea after the earthquake on 12... | [
"Buildings collapsing or showing significant structural damage during the 12 September 2016 earthquake in Gyeongju, South Korea",
"People evacuating their homes and gathering in safe areas in Gyeongju, South Korea after the earthquake on September 12, 2016",
"Emergency responders searching for survivors in rubb... | [
"People running out of buildings during the earthquake in Gyeongju, South Korea on September 12, 2016",
"Rescue teams searching through rubble after the 12 September 2016 earthquake in Gyeongju, South Korea",
"Buildings with cracked walls and fallen debris in Gyeongju, South Korea after the 12 September 2016 ea... | {
"article_url": "https://ko.wikipedia.org/wiki/2016%EB%85%84_%EA%B2%BD%EC%A3%BC_%EC%A7%80%EC%A7%84",
"during_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to decompose the ongoing event in multiple simple events. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"during_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the consequences and aftermath of an earthquake, as these are the most likely to be visualized in a video. Events such as damage to buildings, rescue operations, and people reacting to the quake are plausible and commonly seen in earthquake-related videos.\n\nEVENTS:\n1. People running out of buildings\n2. Rescue teams searching through rubble\n3. Buildings with cracked walls and fallen debris\n4. Emergency vehicles rushing to the scene\n5. People gathering in open spaces for safety",
"en_article_excerpt": "The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.",
"en_article_url": "https://en.wikipedia.org/wiki/2016_Gyeongju_earthquake",
"event_category": "disasters",
"event_name": "gyeongju_earthquake",
"language": "korean",
"prequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what previous events could lead to this given event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"prequel_raw_generation": "EXPLANATION: It makes sense to extract events that are directly related to the aftermath or impact of an earthquake, as these would be the most visually apparent and relevant to a video about the 2016 Gyeongju earthquake. Events such as destruction, evacuation, and rescue efforts are likely to be seen in a video associated with this query.\n\nEVENTS:\n1. People running out of buildings\n2. Emergency responders searching through rubble\n3. Cars driving away from the affected area\n4. Buildings with cracked walls and broken windows\n5. Rescue teams carrying injured people on stretchers",
"sequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what events can be the outcome of this ongoing event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"sequel_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the aftermath and impact of an earthquake, such as damage to infrastructure, rescue efforts, and evacuation procedures. These events are likely to be visually represented in a video associated with the query, providing a concrete and tangible depiction of the earthquake's effects.\n\nEVENTS:\n1. Buildings collapsing or showing significant structural damage\n2. People evacuating their homes and gathering in safe areas\n3. Emergency responders searching for survivors in rubble\n4. Roads blocked by debris and damaged vehicles\n5. Residents assessing damage to their homes and belongings",
"video_URL": "http://youtube.com/watch?v=n6UmQ9Kw2GQ",
"video_description": "경주 지진 이후 여진 307회 발생…\"강도 약해져\" 지난 12일 저녁 발생한 규모 5.8의 경주 지진 탓에 여진이 307차례 발생한 것으로 집계됐습니다. 기상청에 따르면 오늘 오후 4시까지 경주 지진 이후 여진은 규모 1.5∼3.0도 292회 등 모두 307차례 발생했습니다. 기상청 관계자는 \"경주 지진 탓에 여진이 계속 발생하고 있지만 강도가 약해지고 있다\"며 \"앞으로 일어날 여진 규모는 3.0 이하로 예상되기 때문에 그렇게 우려하지 않아도 될 것 같다\"고 말했습니다. 연합뉴스TV : 02-398-4409(제보) 4441(기사문의), 카톡/라인 jebo23 (끝)",
"video_id": "n6UmQ9Kw2GQ"
} | 16 | [
"In the current frame, two news anchors are presenting a weather report. The male anchor is wearing traditional Korean attire and the female anchor is wearing a red and gold patterned hanbok. Behind them is a weather map indicating areas affected by recent weather patterns, with a large portion of the map highlight... | n6UmQ9Kw2GQ | The video depicts a news broadcast covering the aftermath of a 5.8 magnitude earthquake that struck Gyeongju, with a focus on the 307 recorded aftershocks. The news anchors, dressed in traditional Korean attire, present the weather report and discuss the earthquake's impact, displaying a map highlighting areas affected by the recent weather patterns. The video then transitions to surveillance footage and scenes of everyday life, such as a kitchen and a nursery, showing the effects of the aftershocks on household objects. Despite the initial concern, the intensity of the aftershocks is reported to be weakening, with officials stating that there is no need for significant concern. The video concludes with a calm indoor setting, featuring a potted plant and an air conditioning unit, reflecting the diminishing impact of the aftershocks. The news broadcast provides updates on expected travel times on highways, emphasizing the return to normalcy. Overall, the video presents a narrative of the earthquake's aftermath, highlighting the initial disruption and the subsequent return to calm, as the intensity of the aftershocks weakens. | {
"original": " 지난 12일 저녁 발생한 규모 5.8의 경주 지진 탓에 여진이 307차례 발생한 것으로 집계됐습니다. 기상청에 따르면 오늘 오후 4시까지 경주 지진 이후 여진은 규모 1.5에서 3.0도 292회 등 모두 307차례 발생했습니다. 기상청 관계자는 경주 지진 탓에 여진이 계속 발생하고 있지만 강도가 약해지고 있다며 앞으로 일어날 여진 규모는 3.0 이하로 예상되기 때문에 그렇게 우려하지 않아도 될 것 같다고 말했습니다.",
"refined": "On the evening of the 12th, a 5.8 magnitude earthquake occurred in Gyeongju, resulting in 307 aftershocks. According to the weather agency, as of 4 pm today, there have been 307 aftershocks with magnitudes ranging from 1.5 to 3.0 since the Gyeongju earthquake. An agency official stated that although aftershocks are continuing due to the Gyeongju earthquake, their intensity is weakening, and it is expected that future aftershocks will have a magnitude of 3.0 or less, so there is no need to worry.",
"summarized": "A magnitude 5.8 earthquake occurred in Gyeongju on the 12th, resulting in 307 aftershocks, with 292 of them ranging from 1.5 to 3.0 in magnitude. According to the Korea Meteorological Administration, although aftershocks are still occurring due to the earthquake, their intensity is weakening. Officials expect future aftershocks to have a magnitude of 3.0 or less, indicating no significant concern is warranted.",
"translated": "Following the 5.8 magnitude earthquake that struck Gyeongju on the evening of the 12th, a total of 307 aftershocks have been recorded. According to the Korea Meteorological Administration, as of 4 pm today, there have been 307 aftershocks, with 292 of them ranging from 1.5 to 3.0 in magnitude. An official from the Meteorological Administration noted that while aftershocks are still occurring due to the Gyeongju earthquake, their intensity is weakening, and it is expected that future aftershocks will have a magnitude of 3.0 or less, so there is no need for significant concern.",
"translated_llm": "The latest estimates are that 307 earthquakes have occurred due to the 5.8 magnitude earthquake that occurred on the 12th evening. According to the Meteorological Office, as of 4 pm today, the earthquake has caused 307 earthquakes ranging from 1.5 to 3.0 degrees and 292 rounds.",
"translated_whisper": " The last 12 days, there were 5.8 times of earthquakes in Gyeongju, and 307 times of the earthquake in Yeojin. According to the Ministry of Emergency Situations, Yeojin has been 3 times in Gyeongju since the earthquake in Gyeongju, from 1.5 to 3.0 degrees 292 times. The Kisangcheon official said that the earthquake is still happening due to the Gyeongju earthquake, but the intensity is getting weaker, and the expected earthquake size to occur in the future is 3.0 or less, so it is not necessary to worry."
} |
The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea. | [
"People running out of buildings during the 2016 Gyeongju, South Korea earthquake on September 12",
"Emergency responders searching through rubble after the 2016 Gyeongju, South Korea earthquake on September 12, 2016",
"Cars driving away from the affected area in Gyeongju, South Korea after the earthquake on 12... | [
"Buildings collapsing or showing significant structural damage during the 12 September 2016 earthquake in Gyeongju, South Korea",
"People evacuating their homes and gathering in safe areas in Gyeongju, South Korea after the earthquake on September 12, 2016",
"Emergency responders searching for survivors in rubb... | [
"People running out of buildings during the earthquake in Gyeongju, South Korea on September 12, 2016",
"Rescue teams searching through rubble after the 12 September 2016 earthquake in Gyeongju, South Korea",
"Buildings with cracked walls and fallen debris in Gyeongju, South Korea after the 12 September 2016 ea... | {
"article_url": "https://ko.wikipedia.org/wiki/2016%EB%85%84_%EA%B2%BD%EC%A3%BC_%EC%A7%80%EC%A7%84",
"during_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to decompose the ongoing event in multiple simple events. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"during_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the consequences and aftermath of an earthquake, as these are the most likely to be visualized in a video. Events such as damage to buildings, rescue operations, and people reacting to the quake are plausible and commonly seen in earthquake-related videos.\n\nEVENTS:\n1. People running out of buildings\n2. Rescue teams searching through rubble\n3. Buildings with cracked walls and fallen debris\n4. Emergency vehicles rushing to the scene\n5. People gathering in open spaces for safety",
"en_article_excerpt": "The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.",
"en_article_url": "https://en.wikipedia.org/wiki/2016_Gyeongju_earthquake",
"event_category": "disasters",
"event_name": "gyeongju_earthquake",
"language": "korean",
"prequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what previous events could lead to this given event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"prequel_raw_generation": "EXPLANATION: It makes sense to extract events that are directly related to the aftermath or impact of an earthquake, as these would be the most visually apparent and relevant to a video about the 2016 Gyeongju earthquake. Events such as destruction, evacuation, and rescue efforts are likely to be seen in a video associated with this query.\n\nEVENTS:\n1. People running out of buildings\n2. Emergency responders searching through rubble\n3. Cars driving away from the affected area\n4. Buildings with cracked walls and broken windows\n5. Rescue teams carrying injured people on stretchers",
"sequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what events can be the outcome of this ongoing event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"sequel_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the aftermath and impact of an earthquake, such as damage to infrastructure, rescue efforts, and evacuation procedures. These events are likely to be visually represented in a video associated with the query, providing a concrete and tangible depiction of the earthquake's effects.\n\nEVENTS:\n1. Buildings collapsing or showing significant structural damage\n2. People evacuating their homes and gathering in safe areas\n3. Emergency responders searching for survivors in rubble\n4. Roads blocked by debris and damaged vehicles\n5. Residents assessing damage to their homes and belongings",
"video_URL": "http://youtube.com/watch?v=uoagwCVoHGM",
"video_description": "경주 지진으로 8명 부상…재산피해 253건 어제 저녁 경주 일대에서 발생한 지진으로 인해 8명이 다치고 카카오톡과 국민안전처 홈페이지 등 통신망이 일시적으로 마비되는 등의 장애가 있었습니다. 국민안전처는 경주 일대에서 오늘과 어제에 걸쳐 일어난 지진으로 인해 80살 여성 황 모 씨 등 8명이 경상을 입고 253건의 재산피해가 신고됐다고 밝혔습니다. 이번 지진으로 인해 KTX 대전~부산 구간이 5분가량 정차하기도 했습니다. 연합뉴스TV : 02-398-4409(제보) 4441(기사문의), 카톡/라인 jebo23 (끝)",
"video_id": "uoagwCVoHGM"
} | 16 | [
"The image shows a news anchor standing in front of a large screen displaying a black and white surveillance camera footage, presumably related to the earthquake in Gyeongju. The footage appears to depict outdoor scenes, possibly from an urban area, featuring some construction materials and debris. The text on the ... | uoagwCVoHGM | The video depicts the aftermath of an earthquake in Gyeongju, showing significant damage to buildings and infrastructure. The earthquake resulted in 8 injuries, including an 80-year-old woman named Hwang, and 253 cases of property damage. The disaster temporarily disrupted communication networks and caused the KTX train to halt for about 5 minutes. The video shows news anchors reporting on the incident, displaying black and white surveillance footage of the damaged areas, and providing updates on the extent of the damage and the response efforts. The scenes also show people assessing the damage, cleaning up debris, and continuing with their daily activities, such as traveling through Gimhae Airport, which appears to be operating normally despite the earthquake. The overall context is one of disruption and recovery, with a focus on the impact of the earthquake on the community and the efforts to restore normalcy. | {
"original": " 어제저녁 경주 일대에서 발생한 지진으로 인해 8명이 다치고, 카카오톡과 국민안전처 홈페이지 등 통신망이 일시적으로 마비되는 등의 장애가 있었습니다. 국민안전처는 경주 일대에서 오늘과 어제에 걸쳐 일어난 지진으로 인해 80살 여성 황 모 씨 등 8명이 경상을 입고 253건의 재산 피해가 신고됐다고 밝혔습니다. 이번 지진으로 인해 KTX 대전에서 부산 구간이 5분가량 정차하기도 했습니다. MBC 뉴스 김성현입니다.",
"refined": "Yesterday evening, an earthquake in Gyeongju left 8 people injured and caused temporary disruptions to communication networks, including Kakao Talk and the National Safety Agency's website. The National Safety Agency reported that the earthquake, which occurred in Gyeongju over the past two days, resulted in 8 people, including an 80-year-old woman, sustaining minor injuries, and 253 cases of property damage. The earthquake also caused the KTX train service between Daejeon and Busan to be suspended for about 5 minutes.",
"summarized": "An earthquake in Gyeongju resulted in 8 people being injured, including an 80-year-old woman, and 253 cases of property damage reported. The incident also caused a temporary disruption to communication networks, including Kakao Talk and the National Security Agency website, and a 5-minute halt to the KTX train service between Daejeon and Busan.",
"translated": "Yesterday evening, an earthquake in the Gyeongju area resulted in 8 injuries and temporarily disrupted communication networks, including KakaoTalk and the National Security Agency's website. The National Security Agency reported that 8 people, including an 80-year-old woman named Hwang, were injured and 253 cases of property damage were reported due to the earthquakes that occurred in Gyeongju yesterday and today. The earthquake also caused the KTX train to halt for about 5 minutes on the Daejeon-Busan route. This is MBC News reporter Kim Seong-hyeon.",
"translated_llm": "The earthquake in the raceway yesterday evening caused 8 people to be injured, and the telecommunications network including the KaKaoTok and the website of the National Security Agency were temporarily disrupted. The National Security Agency said that eight people including an 80-year-old woman, Mr. Zhuang Mo, were injured and 253 property damage was reported due to the earthquake in the raceway today and yesterday.",
"translated_whisper": " Yesterday evening, 8 people were injured in the earthquake in Gyeongju, and there was a problem with the communication network such as Kakao Talk and the National Security Agency website. The National Security Agency reported that 8 people, including an 80-year-old woman, were injured in the earthquake that occurred yesterday and today in Gyeongju. 253 property damage was reported. Due to the earthquake, the Busan section of KTX Daejeon was also closed for about 5 minutes."
} |
The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea. | [
"People running out of buildings during the 2016 Gyeongju, South Korea earthquake on September 12",
"Emergency responders searching through rubble after the 2016 Gyeongju, South Korea earthquake on September 12, 2016",
"Cars driving away from the affected area in Gyeongju, South Korea after the earthquake on 12... | [
"Buildings collapsing or showing significant structural damage during the 12 September 2016 earthquake in Gyeongju, South Korea",
"People evacuating their homes and gathering in safe areas in Gyeongju, South Korea after the earthquake on September 12, 2016",
"Emergency responders searching for survivors in rubb... | [
"People running out of buildings during the earthquake in Gyeongju, South Korea on September 12, 2016",
"Rescue teams searching through rubble after the 12 September 2016 earthquake in Gyeongju, South Korea",
"Buildings with cracked walls and fallen debris in Gyeongju, South Korea after the 12 September 2016 ea... | {
"article_url": "https://ko.wikipedia.org/wiki/2016%EB%85%84_%EA%B2%BD%EC%A3%BC_%EC%A7%80%EC%A7%84",
"during_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to decompose the ongoing event in multiple simple events. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"during_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the consequences and aftermath of an earthquake, as these are the most likely to be visualized in a video. Events such as damage to buildings, rescue operations, and people reacting to the quake are plausible and commonly seen in earthquake-related videos.\n\nEVENTS:\n1. People running out of buildings\n2. Rescue teams searching through rubble\n3. Buildings with cracked walls and fallen debris\n4. Emergency vehicles rushing to the scene\n5. People gathering in open spaces for safety",
"en_article_excerpt": "The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.",
"en_article_url": "https://en.wikipedia.org/wiki/2016_Gyeongju_earthquake",
"event_category": "disasters",
"event_name": "gyeongju_earthquake",
"language": "korean",
"prequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what previous events could lead to this given event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"prequel_raw_generation": "EXPLANATION: It makes sense to extract events that are directly related to the aftermath or impact of an earthquake, as these would be the most visually apparent and relevant to a video about the 2016 Gyeongju earthquake. Events such as destruction, evacuation, and rescue efforts are likely to be seen in a video associated with this query.\n\nEVENTS:\n1. People running out of buildings\n2. Emergency responders searching through rubble\n3. Cars driving away from the affected area\n4. Buildings with cracked walls and broken windows\n5. Rescue teams carrying injured people on stretchers",
"sequel_prompt": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 26 Jul 2024\n\n<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nYou are given a video search query. Assume the video search query is an ongoing event. You have to find out what events can be the outcome of this ongoing event. The extracted events should be concrete enough to be visualized in a video.\n\nHere are some guidelines: \n- Extract events that could be visualized in a video.\n- Before extracting an event, consider whether the event text could serve as a reasonable caption for a frame of a video associated with the query.\n- Give specific events related to the query, use your previous WORLD KNOWLEDGE to extract events that are likely to be seen in a video. For example, if the query is \"2019 Christmas Eve Protests in Hong Kong\", you might return \"Protesters marching in the streets\", but not \"Protestors marching in snow\" because it is not likely to have snow in Hong Kong, though its pretty common in other places. \n- First, explain your reasoning in 1-2 sentences about which events it makes sense to extract. Then, provide a numbered list with the events (anywhere from one to five event; use your best judgement to determine how many to extract)\n- Structure your response into two sections, \"\"EXPLANATION:\"\" and then \"\"EVENTS:\"\". The events sections should contain nothing but the events. \n\nHere is the query I'd like you to extract events from: The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea.<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
"sequel_raw_generation": "EXPLANATION: Given the query about the 2016 Gyeongju earthquake, it makes sense to extract events that are directly related to the aftermath and impact of an earthquake, such as damage to infrastructure, rescue efforts, and evacuation procedures. These events are likely to be visually represented in a video associated with the query, providing a concrete and tangible depiction of the earthquake's effects.\n\nEVENTS:\n1. Buildings collapsing or showing significant structural damage\n2. People evacuating their homes and gathering in safe areas\n3. Emergency responders searching for survivors in rubble\n4. Roads blocked by debris and damaged vehicles\n5. Residents assessing damage to their homes and belongings",
"video_URL": "http://youtube.com/watch?v=5p1N4MtudiI",
"video_description": "경주 지진 316회 이어가다 현재는 '잠잠' 지난 12일 발생한 규모 5.8의 경주 지진의 여진이 316회 발생한 것으로 집계됐습니다. 기상청은 오늘 오전 11시 기준 여진 횟수가 316회라고 발표했습니다. 규모별 횟수는 1.5에서 3.0이 301차례, 3.0에서 4.0은 14차례, 4.0에서 5.0은 1차례 감지됐습니다. 사흘째 계속된 여진은 일단 멈추는 징후를 보이고 있는 것으로 기상청은 파악하고 있습니다. 연합뉴스TV : 02-398-4409(제보) 4441(기사문의), 카톡/라인 jebo23 (끝)",
"video_id": "5p1N4MtudiI"
} | 16 | [
"In the current frame, two news anchors are presenting on a television broadcast. The male anchor is on the left, and the female anchor is on the right, both dressed in traditional Korean attire. The background shows a cityscape with a river and bridges. The text on the screen reads \"경주 지진 316회 이어가다 현재는 '잠잠'\" whi... | 5p1N4MtudiI | The video is a news broadcast from Korean News TV, focusing on the aftermath of a 5.8-magnitude earthquake that occurred in Gyeongju. The broadcast, which took place at 15:01, reports on the 316 aftershocks that have been recorded, with the majority ranging from 1.5 to 3.0 magnitude. The aftershocks, which have continued for three days, are showing signs of subsiding, according to the Korea Meteorological Administration. The news anchors, dressed in traditional Korean attire, provide updates on the seismic activity, displaying maps and data from various stations, including KIGAM, YBN BHZ KG, DNH BHZ KG, and NSN BHZ KG. The maps and seismic data are used to track the aftershocks, which are represented by colored lines, markers, and dots on the maps, indicating their location, frequency, and intensity. The broadcast also features computer screens displaying seismograms, lists of seismic events, and maps of the Korean Peninsula, highlighting the ongoing monitoring and analysis of the earthquake's impact. The temperature is reported to be around 26-28°C throughout the broadcast. Overall, the video provides a detailed report on the Gyeongju earthquake and its aftershocks, emphasizing the ongoing monitoring and analysis of the seismic activity and its gradual subsidence. | {
"original": " 지난 12일에 발생한 규모 5.8의 경주 지진의 여진이 316회 발생한 것으로 집계됐습니다. 기상청은 오늘 오전 11시 기준 여진 횟수가 316회라고 발표했습니다. 규모별 횟수는 1.5에서 3.0이 301차례, 3.0에서 4.0은 14차례, 4.0에서 5.0은 1차례 감지됐습니다. 사흘째 계속된 여진은 일단 멈추는 징후를 보이고 있는 것으로 기상청은 파악하고 있습니다.",
"refined": "The aftershocks of the 5.8 magnitude Gyeongju earthquake that occurred on the 12th have been recorded 316 times. As of 11 am this morning, the Korea Meteorological Administration reported 316 aftershocks. The breakdown by magnitude is as follows: 301 aftershocks between 1.5 and 3.0, 14 between 3.0 and 4.0, and one between 4.0 and 5.0. The aftershocks, which have continued for three days, are showing signs of subsiding, according to the Korea Meteorological Administration.",
"summarized": "A 5.8-magnitude earthquake occurred in Gyeongju on the 12th, followed by 316 aftershocks as of 11 am today. The aftershocks were mostly minor, with 301 between 1.5 and 3.0 magnitude, 14 between 3.0 and 4.0, and 1 between 4.0 and 5.0. The aftershocks have continued for four days but appear to be subsiding, according to the Korea Meteorological Administration.",
"translated": "The aftershocks of the 5.8-magnitude earthquake that occurred in Gyeongju 12 days ago have been recorded 316 times. As of 11 am today, the Korea Meteorological Administration reported that the number of aftershocks has reached 316. By magnitude, 301 of the aftershocks ranged from 1.5 to 3.0, 14 ranged from 3.0 to 4.0, and one ranged from 4.0 to 5.0. The aftershocks, which have continued for three days, are showing signs of subsiding, according to the Korea Meteorological Administration.",
"translated_llm": "The number of earthquakes in the last 12 days has been estimated at 316 of magnitude 5.8. The Meteorological Office has announced that the number of earthquakes as of 11 am today is 316. The number of earthquakes in the past 12 days has been estimated at 1.5 to 3.0 and 301 times, 3.0 to 4.0 and 14 times, 4.0 to 5.0 and once.",
"translated_whisper": " The incident occurred on 5.8 Gyeongju. Three hundred and sixteen times of earthquake occurred. As a result. The Kisangcheon News Agency reported that the number of earthquakes occurred at 11 a.m. this morning as 316. The number of a 1.5 to 3.0 is 301 times. Three three points zero four points zero were a few zeroes, four points zero and a quarter is a few. I have been detected. Over the past four days, the 여진 has been seen to be stuck for a while. Kisang-chung is knowing that."
} |
The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea. | ["People running out of buildings during the 2016 Gyeongju, South Korea earthquake on September 12",(...TRUNCATED) | ["Buildings collapsing or showing significant structural damage during the 12 September 2016 earthqu(...TRUNCATED) | ["People running out of buildings during the earthquake in Gyeongju, South Korea on September 12, 20(...TRUNCATED) | {"article_url":"https://ko.wikipedia.org/wiki/2016%EB%85%84_%EA%B2%BD%EC%A3%BC_%EC%A7%80%EC%A7%84","(...TRUNCATED) | 16 | ["The image depicts a traditional Korean village with numerous rooftops made of dark gray tiles. In (...TRUNCATED) | roaxxb8zRLU | "A traditional Korean village, specifically in the Gyeongju City, Hwangnam-dong area, is depicted in(...TRUNCATED) | {"original":" ល្ពបានដោយើងមំខងានដំល ១ េការអណា(...TRUNCATED) |
The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea. | ["People running out of buildings during the 2016 Gyeongju, South Korea earthquake on September 12",(...TRUNCATED) | ["Buildings collapsing or showing significant structural damage during the 12 September 2016 earthqu(...TRUNCATED) | ["People running out of buildings during the earthquake in Gyeongju, South Korea on September 12, 20(...TRUNCATED) | {"article_url":"https://ko.wikipedia.org/wiki/2016%EB%85%84_%EA%B2%BD%EC%A3%BC_%EC%A7%80%EC%A7%84","(...TRUNCATED) | 16 | ["The current frame shows a news broadcast on a South Korean channel, featuring two news anchors rep(...TRUNCATED) | 3ojHTwrvx90 | "The video depicts the government's response to a series of strong earthquakes in Gyeongju, Gyeongbu(...TRUNCATED) | {"original":" 박근혜 대통령이 최근 강진이 잇따라 발생한 경북 경주지역을 찾(...TRUNCATED) |
The 2016 Gyeongju earthquake occurred on September 12 near Gyeongju, South Korea. | ["People running out of buildings during the 2016 Gyeongju, South Korea earthquake on September 12",(...TRUNCATED) | ["Buildings collapsing or showing significant structural damage during the 12 September 2016 earthqu(...TRUNCATED) | ["People running out of buildings during the earthquake in Gyeongju, South Korea on September 12, 20(...TRUNCATED) | {"article_url":"https://ko.wikipedia.org/wiki/2016%EB%85%84_%EA%B2%BD%EC%A3%BC_%EC%A7%80%EC%A7%84","(...TRUNCATED) | 16 | ["The current frame features a news anchor dressed in traditional Korean attire, likely presenting a(...TRUNCATED) | w6jdwGYcKZU | "The video depicts the aftermath of a 5.8 magnitude earthquake in Gyeongju, showing significant prop(...TRUNCATED) | {"original":" 국민안전처는 경주에서 발생한 규모 5.8의 지진으로 부상자 23명(...TRUNCATED) |
End of preview. Expand
in Data Studio
README.md exists but content is empty.
- Downloads last month
- 6
Size of downloaded dataset files:
18 MB
Size of the auto-converted Parquet files:
18 MB
Number of rows:
2,394