일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | ||
6 | 7 | 8 | 9 | 10 | 11 | 12 |
13 | 14 | 15 | 16 | 17 | 18 | 19 |
20 | 21 | 22 | 23 | 24 | 25 | 26 |
27 | 28 | 29 | 30 |
- 데이터 시각화
- GPT
- 현대자동차
- 데이터 구축
- Data Viz
- RNN
- pyTorch
- AI Math
- Transformer
- Ai
- mrc
- 2023 현대차·기아 CTO AI 경진대회
- passage retrieval
- Self-attention
- Bart
- 기아
- KLUE
- 딥러닝
- N21
- Bert
- word2vec
- Attention
- Optimization
- matplotlib
- AI 경진대회
- nlp
- dataset
- seaborn
- ODQA
- N2N
- Today
- Total
목록passage retrieval (2)
쉬엄쉬엄블로그

이 색깔은 주석이라 무시하셔도 됩니다. Passage Retrieval-Dense Embedding Dense Embedding Introduction to Dense Embedding Passage Embedding 구절(Passage)을 벡터로 변환하는 것 Sparse Embedding TF-IDF 벡터는 Sparse하다. 0이 0이 아닌 값보다 많다 Limitations of sparse embedding 차원의 수가 매우 크다 → compressed format으로 극복 가능 유사성을 고려하지 못한다 Dense Embedding이란? Complementray to sparse representations by design 더 작은 차원의 고밀도 벡터 (length = 50-1000) 각 차원이 특..

이 색깔은 주석이라 무시하셔도 됩니다. Passage Retrieval-Sparse Embedding Sparse Embedding Introduction to Passage Retrieval Passage Retrieval 질문(query)에 맞는 문서(passage)를 찾는 것 Passage Retrieval with MRC Open-domain Question Answering : 대규모의 문서 중에서 질문에 대한 답을 찾기 Passage Retrieval과 MRC를 이어서 2-Stage로 만들 수 있음 Overview of Passage Retrieval Query와 Passage를 임베딩한 뒤 유사도로 랭킹을 매기고, 유사도가 가장 높은 Passage를 선택함 Passage Embedding a..