DL researcher: we lack a theory of deep learning because you can't hill climb on it
jasondeanlee · x · 2026-11-21
jasondeanlee says he'd like to see a real theory of deep learning, but argues researchers "can't hill climb on that yet" — theoretical work can't currently be iteratively optimized the way empirical research can.
More from Research
- OpenAI's GPT-Rosalind agents surface new drug discovery lead for neurodegenerative disease — OpenAI · 2026-11-21
- Vik Sikka on the reported Claude-assisted proof of the Riemann Hypothesis: 'an incredible day' — vsikka · 2026-11-21
- MIT's Sutherland Verifies Riemann Hypothesis Lean Proof: All 175,187 Modules Compile — altryne · 2026-11-21
- Live Proof Monitor Launched for Independent Lean Verification of Riemann Hypothesis Proof — altryne · 2026-11-21
- New proof shows finite-time blow-up for unforced Navier–Stokes, with Lean 4 formalization — IgorCarron · 2026-11-21
- Independent mathematicians attest AI Riemann hypothesis proof is almost certainly true — soumitrashukla9 · 2026-11-21