CTWM beats 18M-param LeWM on six goal-reaching tasks with half the parameters (9M)
hisspikeness · x · 2026-10-05
The author presents CTWM results on six pixel-based goal-reaching tasks (navigation and manipulation): it matches or beats LeWM, a strong task-agnostic baseline, using half the parameters (9M vs. 18M). On PointMaze, CTWM's latent distances track the maze's true commute-time structure far more closely than LeWM, especially around bottlenecks — exactly where it matters most for planning.
More from Research
- Schmidhuber: Linearized self-attention Transformers date back to his 1992 NECO paper — SchmidhuberAI · 2026-10-05
- ML Conference AC warns: papers that are 'incomprehensible' will be desk rejected — tyrell_turing · 2026-10-05
- Meta's ProWAM hits 70% zero-shot real-world robot success with sparse visual sub-goals — meta · 2026-10-05
- Interpretability Researcher: Don't Let Unexplained Model Behavior Block You — repligate · 2026-10-05
- Yandex Music replaced 15+ candidate generators with one transformer, +6.3% listening time — SettingAccording8986 · 2026-10-05
- Cohere Labs lands multiple papers and a workshop talk at COLM — Cohere_Labs · 2026-10-05