Mmastodon TechnologyTechnology first seen 1 d ago, last 1 d ago, peak #4
Dust: Pretraining Transformers Without Backpropagation
Original: Dust: Pretraining Transformers Without Backpropagation https://qlabs.sh/research/dust # HackerNews # Tech # AI
A research project called Dust claims a method for pretraining transformer models without backpropagation, the algorithm at the core of modern deep learning. If the results hold up, the approach could challenge assumptions about how large models must be trained, but independent verification and details of its performance are not yet established.
Why now: Removing backpropagation from transformer training would be a significant shift in AI research, drawing attention from practitioners.
Rank over time, top of the chart is #1. 2 snapshots from 1 d ago to 1 d ago.
Evidence
- Dust: Pretraining Transformers Without Backpropagation https://qlabs.sh/research/dust # HackerNews # Tech # AI · hackernews@robot.villas · 3
API: https://socialmediatrends-api.osmike.com/v1/trends/1172100