Yi Ma Reaffirms Closed-Loop Learning: End-to-End Models Will Return to Feedback
YiMaTweets · x · 2026-08-06
UC Berkeley professor Yi Ma reflects on the evolution of AI architectures over the past few years. He notes that while the industry enthusiastically celebrated end-to-end models, his team made the controversial argument that all truly intelligent systems must learn via closed-loop feedback (detailed in their paper 'CTRL: closed-loop transcription'). Although the paper struggled to get published in main ML venues at the time, Ma believes the industry is eventually making a full circle back to this correct approach, predicting it will soon become mainstream.
Related event: Ma Yi Reiterates End-to-End Models Will Return to Closed-Loop Learning(3 posts)→
More from Research
- Berkeley and DeepMind Propose Instruct-to-Act: Making World-Model Controllers Language-Instructable — berkeley_ai · 2026-09-22
- Why OpenAI bets on math: it's the most verifiable domain for reinforcement learning — burny_tech · 2026-09-22
- Must-read papers of the week: recursive self-improvement, world models, KV cache compression — TheTuringPost · 2026-09-22
- Meta's A-MLE agent automates ML experimentation for ads ranking, cutting error 2.56% — rohanpaul_ai · 2026-09-22
- Blind RSA apps like Privacy Pass face real-world threat model from scaled oracle queries — matthew_d_green · 2026-09-22
- Harvard/MIT paper FINSKILLOPS makes financial AI self-improve via regression-tested skills — rohanpaul_ai · 2026-09-22