AI Paper Atlas

Research map / Large language models / Agents, code and math

Rethinking Visual Intelligence: Insights from Video Pretraining

Pablo Acuaviva, Aram Davtyan, Mariam Hassan, Sebastian Stapf, Ahmad Rahimi, Alexandre Alahi, Paolo Favaro

ICML 2026 · Large language models › Agents, code and math · approach: Mathematical and structural reasoning benchmarks

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use a pretrained video diffusion model instead of an LLM as the backboneLoRA-adapted CogVideoX1.5-5B vs Qwen3-4B-Instruct-2507 · Route planning (Shortest Path, Maze)training samples needed: up to a tenfold reduction in data requirements in low-sample regimesImproved
use a video diffusion backbone instead of an LLMCogVideoX1.5-5B · Visual games (Hitori, Sudoku, Sudoku Mini, Connect 4)accuracy vs training set sizeImproved
use a video diffusion backbone instead of an LLMCogVideoX1.5-5B · Chess Mate-in-1accuracy vs training set sizeHurt
train on 13x13 mazes and test on 21x21 with a video diffusion backboneCogVideoX1.5-5B · Maze GeneralizationaccuracyImproved
use a video diffusion backbone instead of an LLMCogVideoX1.5-5B · 1D elementary cellular automata (delta 0.9)examples to reach 0.9 accuracyNo clear effect
use a video diffusion backbone instead of an LLMCogVideoX1.5-5B · Life-like cellular automataexamples to reach 0.9 accuracyImproved
use a video diffusion backbone, with longer prediction horizonCogVideoX1.5-5B · Langton's Ant (horizon 2, 3, 5, 10)accuracy vs training set sizeImproved
raise LoRA rank from 32 to 64 for Qwen3 (and lower 64 to 32 for CogVideoX) (ablation)Qwen3 and CogVideoX1.5-5B LoRA · ARC and structured tasksaccuracy: rank 32 vs rank 64No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy