AssistGPT is a multi-modal AI assistant that integrates LLMs with various tools using a Plan, Execute, Inspect, and Learn (PEIL) approach to address complex visual-based tasks. It achieved state-of-the-art results on A-OKVQA and NExT-QA benchmarks.

https://arxiv.org/abs//2306.08640

YouTube: https://www.youtube.com/@ArxivPapers

PODCASTS:

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers