Hacker News new | ask | show | jobs
HippoRAG: Neurobiologically Inspired Long-Term Memory for LLMs (2024) (arxiv.org)
65 points by veryluckyxyz 505 days ago
3 comments

I am not a biologist, but the first sentence seems bold to me:

> Millions of years of evolution have led mammalian brains to develop the crucial ability to store large amounts of world knowledge and continuously integrate new experiences without losing previous ones.

My impression has always been that humans have been good at selective forgetting, hence keeping relevant memories and dropping others.

Edit: it looks like none of the authors has a biological background either. How serious do they mean the "neurobiologically inspired" claim?

It's weird that's the first sentence of the abstract is different:

> In order to thrive in hostile and ever-changing natural environments, mammalian brains evolved to store large amounts of knowledge about the world and continually integrate new information while avoiding catastrophic forgetting.

I think that's more aligned with selective forgetting

> HippoRAG [..] builds a KG from scratch using LLMs and performs multi-hop retrieval without any supervision

They use GPT-3.5 Turbo[0] for entity extraction when populating the knowledge graph.

Curious how well GPT-4/Claude would do here?

[0]https://arxiv.org/html/2405.14831v3#A3

This was released on 23 May 2024, and the source code is on github[0]. It didn't pickup traction like other RAG approaches. I haven't tested it so i don't have an opinion on it.

[0] https://github.com/OSU-NLP-Group/HippoRAG