AI.info
Jan Kautz
Explore Jan Kautz on AI.info.
- OSWorld-Pro: Process-based Evaluation for Computer Use Agents
- Attend Before Attention: Efficient and Scalable Video Understanding via Autoregressive Gazing
- World Action Models are Zero-shot Policies
- Learning to Discover at Test Time
- Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent Planning
- NVIDIA Nemotron 3: Efficient and Open Intelligence
- Efficient-DLM: From Autoregressive to Diffusion Language Models, and Beyond in Speed
- GSPN-2: Efficient Parallel Sequence Modeling
- ToolOrchestra: Elevating Intelligence via Efficient Model and Tool Orchestration
- Nemotron-Flash: Towards Latency-Optimal Hybrid Small Language Models
- SONIC: Supersizing Motion Tracking for Natural Humanoid Whole-Body Control
- World Simulation with Video Foundation Models for Physical AI
- ProfBench: Multi-Domain Rubrics requiring Professional Knowledge to Answer and Judge
- OmniVinci: Enhancing Architecture and Data for Omni-Modal Understanding LLM