REEL: Learning to Retrieve, Evaluate, and Evolve Visual Memory for Any Long-Horizon Multi-Shot Video Generation
Memory-based long-video generation aims to combine user control with long-range visual consistency.
ProofPaper ↗
Key points
- Representative multi-shot systems couple memory to generator-specific interfaces or use predefined agent workflows, rather than independently training a memory policy for cross-shot objectives.
- We introduce REEL, a general-purpose visual-memory planner.
- To train REEL, we construct two high-quality datasets, REEL-SFT-4k and REEL-RL-1k, covering diverse interaction patterns and prompt types.
- REEL-Bench tests identity preservation and intended state changes under text-only requests and timed reference-image uploads.
Sources (1)
- [1]REEL: Learning to Retrieve, Evaluate, and Evolve Visual Memory for Any Long-Horizon Multi-Shot Video GenerationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 12, 11:54 AM
Memory-based long-video generation aims to combine user control with long-range visual consistency.
Representative multi-shot systems couple memory to generator-specific interfaces or use predefined agent workflows, rather than independently training a memory policy for cross-shot objectives.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 12, 2026AUGUstus: Autonomous Driving VLA Model with Unified Generation and Understanding
- Oct 12, 2026InfoVis: Unified Slide Image Generation and Editing
- Oct 11, 2026RSIPO: Recursive Self-Improving Policy Optimization for Agentic Reinforcement Learning via Skill-Policy Co-Evolution
- Oct 11, 2026Data-Centric Recursive Self-Improvement for Navigation via Policy-Synchronized Language-Action Dual Bootstrapping
- Oct 10, 2026KV4-RL: Unlocking Faster Asynchronous Rollouts for RLVR with INT4 KV Cache Quantization
- Oct 10, 2026LightVAE: Towards Compact and Efficient Video Autoencoders