#100 Dr. PATRICK LEWIS - Retrieval Augmented Generation

Dr. Patrick Lewis is a London-based AI and Natural Language Processing Research Scientist, working at co:here. Prior to this, Patrick worked as a research scientist at the Fundamental AI Research Lab (FAIR) at Meta AI. During his PhD, Patrick split his time between FAIR and University College London, working with Sebastian Riedel and Pontus Stenetorp.

Patrick’s research focuses on the intersection of information retrieval techniques (IR) and large language models (LLMs). He has done extensive work on Retrieval-Augmented Language Models. His current focus is on building more powerful, efficient, robust, and update-able models that can perform well on a wide range of NLP tasks, but also excel on knowledge-intensive NLP tasks such as Question Answering and Fact Checking.

References:

Patrick Lewis (Natural Language Processing Research Scientist @ co:here)

Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks (Patrick Lewis et al)

Atlas: Few-shot Learning with Retrieval Augmented Language Models (Gautier Izacard, Patrick Lewis, et al)

Improving language models by retrieving from trillions of tokens (RETRO) (Sebastian Borgeaud et al)

[grayboywilliams]

RAG is super interesting with so much potential. Cant believe this videos been here since February.

[loveolutionaustin]

Great guest! Thanks for highlighting this work! 👏😸

[earleyelisha]

I smiled once when I saw the MLST notification and then a second time within the first 5 seconds! 2 ✌🏾points Tim!!

[rick-kvgl]

thanks for giving them som
e love, really interesting startup in terms of hacking their way vs brute forcing

[AnnieCushing]

This was infinitely fascinating. I was asking ChatGPT for explanations all throughout.

[Dima-rjbv]

IMHO highly underestimated episode. With all the hype surrounding RAG and LLMs and hallucinations, this short interview gives a great mental framework. ❤

[mittanuljak]

// next time please use a different camera placement, where the microphone is not in the face of the interviewee. + its a little bit strange to see the reporter on the divided screen all time. usually it is important if his reaction, face expression has some meaning in the interview.

[GBlunted]

RAG is so over rated... right? 🤔 Why is it given so much attention, I don't get it. Until it can actually answer questions or derive information that isn't explicitly found in the document store it seems kind of weak... Like if you encode Alice in Wonderland, what kind of answer am I going to get if I ask, "What's the symbolic significance of Alice's repeated changes in size throughout the story?"

Or if I vectorize and index all the Disney movies and ask, "How do the evolution of the themes of 'family' and 'friendship' reflect the changing societal values and expectations over the years these Disney films were released?"

I mean, RAG isn't going to be any help here right?

[DeepakSharma-brfw]

Thats a very disrespectful way to talk to your guest. I guess with lot of views comes lot of pride and hubris.