AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Atlas: Few-shot Learning with Retrieval Augmented Language Models

Gautier Izacard, Patrick Lewis, Maria Lomeli, Lucas Hosseini, Fabio Petroni, Timo Schick, Jane Dwivedi-Yu, Armand Joulin, Sebastian Riedel, Edouard Grave

JMLR 2023 · Large language models › Retrieval and knowledge · approach: Knowledge-augmented question answering · 744 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add joint pre-training of retriever and reader (ablation)Atlas · KILT 64-shot (NQ/WoW/FEVER)average: 30.0 -> 42.2Improved
train the retriever during pre-training (Perplexity Distillation) instead of fixed (ablation)Atlas · KILT 64-shot avgaverage: 42.2 -> 45.7Improved
train the retriever during pre-training instead of fixed (ablation)Atlas · KILT 1024-shot avgaverage: 51.1 -> 51.0No clear effect
use masked language modelling instead of prefix language modelling pretext task (ablation)Atlas · KILT 64-shot avgaverage: 40.1 -> 42.4Improved
use Wikipedia index instead of Common Crawl index with Wikipedia training data (ablation)Atlas · KILT 64-shot avgaverage: 39.8 -> 42.4Improved
fine-tune retriever query-side only instead of fixed retriever (ablation)Atlas · KILT 64-shot avgaverage: 41.1 -> 45.7Improved
full fine-tuning instead of query-side fine-tuning (ablation)Atlas · KILT 1024-shot NQexact match: 44.9 -> 47.0Improved
top-100 re-ranking instead of full index update (ablation)Atlas · KILT 64-shot avgaverage: 44.1 -> 44.7No clear effect
use de-biased cyclic-permutation inference (ablation)Atlas-11B · MMLU zero-shotaccuracy: 36.8 -> 47.1Improved
use de-biased cyclic-permutation inference (ablation)Atlas-11B · MMLU 5-shot (multi-task)accuracy: 56.4 -> 56.6No clear effect
use retrieval (Atlas) instead of closed-book T5T5 · MMLU 5-shot, 11Baccuracy: 36.1 -> 43.4Improved
use temporally matched Wikipedia index (ablation)Atlas · NaturalQuestions 64-shotexact match: 41.3 -> 45.1Improved

Later papers that compared with this method

3 later accepted papers report Atlas base as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy