true-jepa hits 89.7% unsupervised for new SOTA, author targets 100%
MarcJSchmidt · x · 2026-11-22
Marc Schmidt reports progress on true-jepa, a JEPA-style self-supervised variant: on a new SSLvMB benchmark run the new model scores 83.8%, while true-jepa xD reaches 89.7%, a new SOTA, and another version is at 94.5% — all fully unsupervised and still undertrained.
Responding to a question, he says he's confident of reaching 100%, and that the approach may be general enough to work on pixels as a world model or policy for tasks like Push-T — though that remains unverified.
More from Research
- MIT study: coding agents boost commits 240% but releases only 30% — apps ship into the void — GaryMarcus · 2026-11-22
- ISO: Fixed-Spectrum Optimization Merges RL Experts on 375B MoE in Minutes — zdhnarsil · 2026-11-22
- Redditor trains Krea2 to skip VAE Decode and output pixels directly using DWD paper — Remote_Net374 · 2026-11-22
- 'AI proves the Riemann hypothesis in 2026' — versus Riemann's 20-watt brain in 1859 — doodlestein · 2026-11-22
- Scientific American's 'End of Math' cover resurfaces after Riemann hypothesis reportedly solved — EricTopol · 2026-11-22
- 'Don't train on model outputs' is unenforceable when the data is all over the open web — anpaure · 2026-11-22