Multics Scales Four-Player Training to Ten-Player Inference

Featured

• Multics: Generative Multi-Player World Model with Hierarchical Communication score 10
  入选 HF Daily Papers; HF 热度: 55 upvotes (+4); 有代码实现; 关键词(1): scaling
• AReA: Adaptive Receptive-Field Attention for Efficient High-Resolution Diffusion Transformers score 9
  入选 HF Daily Papers; HF 热度: 42 upvotes (+4); 有代码实现

Also Worth Noting

• MaskSense: Confronting the Visual Exploration Trap in Masked Image Generation score 4
  关键词(2): GRPO, text-to-image; 顶会接收: NeurIPS
• VL-DocIR: A Benchmark for Vision-Based Long Document Information Retrieval score 4
  关键词(1): leaderboard; 顶会接收: NeurIPS
• Compute-optimal data scaling for neural surrogates via multi-fidelity training score 4
  关键词(2): scaling, synthetic data; 顶会接收: NeurIPS
• MAGNET: Manifold-Aware Graph Diffusion Network for Connectome Generation score 3
  顶会接收: NeurIPS
• When Goals Leak Across Relationships: How Cached Goal Representations Shape Cross-Client Decisions score 3
  顶会接收: NeurIPS