Dr. Patrick Lewis
VERIFIED TECHNICAL DOSSIERSources checked

Dr. Patrick Lewis

Cohere

2 Verified ArtifactsSource Checked & Attributed

Verified Proof of Work Artifacts

2 items cataloged

Each artifact below represents an authenticated research publication, production code repository, or technical architectural framework directly authored or co-created by Dr. Patrick Lewis. Every entry undergoes editorial source verification.

#1
RESEARCH Checked Sep 21, 2026

KILT: Benchmarking Knowledge Intensive Language Tasks for Grounded Generation

Developed the KILT benchmark standardizing evaluation across open-domain QA, fact checking, entity linking, and dialogue, measuring both answer generation accuracy and provenance attribution.

Model & Execution Context:Unified Wikipedia corpus snapshot with passage-level attribution targets.
Scope & Limitations

Corpus snapshot drift requires periodic index refreshes to keep knowledge-intensive models updated with real-world facts.

#2
RESEARCH Checked Sep 21, 2026

Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks

Authored the foundational NeurIPS 2020 paper defining Retrieval-Augmented Generation (RAG), introducing end-to-end differentiable architectures that ground generative models against external non-parametric neural indices.

Model & Execution Context:DPR (Dense Passage Retriever) coupled with BART sequence-to-sequence generation over Wikipedia index.
Scope & Limitations

Early RAG formulated single-step retrieval; complex multi-hop reasoning requires iterative or agentic retrieval loops.