Seonglae ChoOpen-domain QA
QA from large Text Corpus
Wikipedia 같은 document 기반 retrieval 해서 QA
기본적으로 top-k올리면 성능 다 올라감
ODQA Frameworks
Exact match or f1 might not good way to evaluation
How to Build an Open-Domain Question Answering System?
[Updated on 2020-11-12: add an example on closed-book factual QA using OpenAI API (beta). A model that can answer any question with regard to factual knowledge can lead to many useful and practical applications, such as working as a chatbot or an AI assistant🤖. In this post, we will review several common approaches for building such an open-domain question answering system. Disclaimers given so many papers in the wild: Assume we have access to a powerful pretrained language model.
https://lilianweng.github.io/posts/2020-10-29-odqa/
오픈도메인 QA 리서치: Open Domain Question Answering
Open-domain question answering : 다양한 주제에 대한 대량의 문서 집합으로부터 자연어 질의에 대한 답변을 찾아오는 태스크 DATA & TASKs [ Natural Questions ] ✅ 구글에 입력된 real query에 대해 long / short / others 타입의 QA - Open-domain QA 테스트를 위해 질문만 취하고, 답변을 찾을 수 있는 문단 정보는 삭제하는 방식으로 실험 진행 - long answer type의 경우 extractive snippet이라고 판단, 제거하고 실험 - (예. 답변이 5토큰 이내인 질문에 대해서만 실험, Lee et al., 2019) - 링크: ai.google.com/research/NaturalQuestions/ [ Cur..
https://littlefoxdiary.tistory.com/76
Papers with Code - Open-Domain Question Answering
Open-domain question answering is the task of question answering on open-domain datasets such as Wikipedia.
https://paperswithcode.com/task/open-domain-question-answering
Contextual compression | 🦜️🔗 Langchain
One challenge with retrieval is that usually you don't know the specific queries your document storage system will face when you ingest data into the system. This means that the information most relevant to a query may be buried in a document with a lot of irrelevant text. Passing that full document through your application can lead to more expensive LLM calls and poorer responses.
https://python.langchain.com/docs/modules/data_connection/retrievers/contextual_compression/
