<Retrieval> [GenRead] Generate rather than Retrieval: Large Language Models are Strong Context Generators (2023.01)

<Retrieval> [GenRead] Generate rather than Retrieval: Large Language Models are Strong Context Generators (2023.01)

관심있는 NLP 논문을 읽어보고 간단히 정리했습니다. (Language & Knowledge Lab의 Retreival 관련) 혹시 부족하거나 잘못된 내용이 있다면 댓글 부탁드립니다 ️ [Microsoft Cognitive Service Research], ICLR 2023 conference paper - GenRead, LLM이 question을 기반으로 contextual document를 생성하고, 최종 답변을 생성하기 위해 해당 document를 참조하는 방식 - 외부 knowledge source로부터 어떤 문서도 retrieve하지 않고 여러 태스크에서 SoTA 성능 달성 - retrieval & generation 형태로 결합도 가능 배경 knowledge-intensive task를..

원문링크 : <Retrieval> [GenRead] Generate rather than Retrieval: Large Language Models are Strong Context Generators (2023.01)

 등록된 다른 글

Object Localization

Object Localization

Neural Style Transfer(5) : Style Cost Function

Neural Style Transfer(5) : Style Cost Function

<LM> DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing

<LM> DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing

Optimization Algorithms(1) - Mini-batch

Optimization Algorithms(1) - Mini-batch

[BOJ] 11404 : 플로이드 [그래프이론](Python)

[BOJ] 11404 : 플로이드 [그래프이론](Python)

<Benchmark, CoT> [Google, REVEAL] A Chain-of-Thought Is as Strong as Its Weakest Link: A Benchmark for Verifiers of Reasoning Chains (2024.02)

<Benchmark, CoT> [Google, REVEAL] A Chain-of-Thought Is as Strong as Its Weakest Link: A Benchmark for Verifiers of Reasoning Chains (2024.02)

TOEIC/TEPS 동시 준비 2주 만에 955/409점 달성한 후기

TOEIC/TEPS 동시 준비 2주 만에 955/409점 달성한 후기

[BOJ] 1916 : 최소비용 구하기 [그래프이론](Python)

[BOJ] 1916 : 최소비용 구하기 [그래프이론](Python)