Information Extractionh
Provide the largest Chinese information extraction dataset in the industry, making machines gain the ability to automatically acquire knowledge from massive natural language texts.
Background
About the Task
1. Task Description
2. Dataset
index
subject_type
predicate
object_type
example
1LocationaltitudeNumber{"object_type": "Number", "predicate": "海拔", "object": "2,240米", "subject_type": "地点", "subject": "卡萨布兰卡火山"}
2TVShowguestPerson{"object_type": "人物", "predicate": "嘉宾", "object": "黄小琥", "subject_type": "电视综艺", "subject": "全能星战"}
3TVShowhostPerson{"object_type": "人物", "predicate": "主持人", "object": "撒贝宁", "subject_type": "电视综艺", "subject": "梦想星搭档"}
4SongsingerPerson{"object_type": "人物", "predicate": "歌手", "object": "李克勤", "subject_type": "歌曲", "subject": "爱不释手"}
5SongfromAlbumAlbum{"object_type": "音乐专辑", "predicate": "所属专辑", "object": "爱不释手 新城唱好音乐大派对", "subject_type": "歌曲", "subject": "爱不释手"}
6SonglyricsPerson{"object_type": "人物", "predicate": "作词", "object": "林夕", "subject_type": "歌曲", "subject": "爱不释手"}
7SongcomposerPerson{"object_type": "人物", "predicate": "作曲", "object": "陈辉阳", "subject_type": "歌曲", "subject": "爱不释手"}
8CountrylanguageLanguage{"object_type": "语言", "predicate": "官方语言", "object": "意大利语", "subject_type": "国家", "subject": "意大利"}
9CountrycapitalCity{"object_type": "城市", "predicate": "首都", "object": "羊苴咩城", "subject_type": "国家", "subject": "大理国"}
10Administrative areaareaNumber{"object_type": "Number", "predicate": "面积", "object": "188平方公里", "subject_type": "行政区", "subject": "河西镇"}
3. Evaluation Metrics
4. Baseline Systems
Participation Info
1. Eligibility
2. Registration
3. Registration Deadline
Scan QR code to join group chat
Timeline
Feb 25
Registration open, training and dev set available
Mar 31
Registration close, partial test set available
May 13
Full test set available
May 20
Testing results submission due
May 31
Final results announcement, system report submission
Aug 24
Workshop and award ceremony
Awards Setting
¥30,000
award certification
¥20,000
award certification
¥10,000
award certification