Retrieval-Augmented Generation

التوليد المعزَّز بالاسترجاع

بنية هجينة تُعزِّز النموذج اللغوي التوليدي بنظام استرجاع خارجي يجلب المستندات ذات الصلة من قاعدة معرفية قبل التوليد، مما يؤسّس الإجابات على أدلة فعلية بدلاً من الاعتماد حصرياً على معاملات محفوظة.

A hybrid architecture that augments a generative language model with an external retrieval system that fetches relevant documents from a knowledge base before generation, grounding answers in actual evidence rather than relying solely on memorized parameters.

Also translated asالتوليد المدعوم بالاسترجاع، RAG، التوليد القائم على الاسترجاع

First appears in this corpus in: Okapi at TREC-3 (1994)

Appears in these papers

Atlas: Few-shot Learning with Retrieval Augmented Language Models2023 in the sky ✦
Atlas: Few-shot Learning with Retrieval Augmented Language Models2023 in the sky ✦
Okapi at TREC-31994 in the sky ✦
Billion-Scale Similarity Search with GPUs2017 in the sky ✦
Fusion-in-Decoder: Leveraging Passage Retrieval with Generative Models for Open Domain Question Answering2021 in the sky ✦
Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks2020 in the sky ✦
REALM: Retrieval-Augmented Language Model Pre-Training2020 in the sky ✦
RETRO: Improving Language Models by Retrieving from Trillions of Tokens2022 in the sky ✦
Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection2023 in the sky ✦
Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection2023 in the sky ✦
WebGPT: Browser-Assisted Question-Answering with Human Feedback2021 in the sky ✦