FrontierPhysics
On FrontierPhysics, no AI configuration among 26 passes more than 46 of 168 end-to-end physics research rollouts.
Evaluating Agents for End-to-End Physics Research
AI Research Radar
Thursday, November 5: 2 frontier papers out of 326 read.
Frontier
Papers rated as moving the field
How we chooseBEAKER
“Results show that current models can follow coarse experimental procedures in videos, but still struggle to ground actions into operable regions and feasible trajectories, to recognize fine-grained laboratory apparatus, states, and safety norms, and to order steps in mid-length procedures.”
This week in research
Oct 30 – Nov 5: 319 cleared, 22 frontier
- Security106◆ 5
- Agents46◆ 4
- Safety40◆ 4
- Society31
- Robotics29◆ 3
- Vision17◆ 2
- 4 moreTraining, Science, ML, Reasoning17◆ 3
- Inference14
- Language10
- Security106◆ 5
- 8 moreRobotics, Vision, Inference, Language, Training, Science, ML, Reasoning87◆ 9
- Agents46◆ 4
- Safety40◆ 4
- Society31
Size: cleared papers◆ and tint: frontier papers9 papers not yet sorted into a field
Models under test
Papers this week that name the model
Labs publishing
Papers this week with an author at the lab
Also cleared
29 papers, by field
Agents7
- From Handcrafted to Generative: RLVR-Optimized Icon Generation for Alexa+Amazon
- Agentic Behavior Engineering: A Design-by-Contract Methodology for Specifying, Verifying, and Governing Enterprise AI Agents
- Anduin: A sybil-resistant agentic marketplace reputation system
- ViBench 2.0: A Benchmark for Long-Horizon Software Building
- The Harness Is the Model: Agent Scaffolding Changes a Fixed LLM's Compliance with Banking Conduct Rules, Unevenly Across Models
Security & privacy7
- Timing is Everything: Inferring Temporal Membership in Continually Updated Models
- Enforcing Accountable Disclosure in Confidential AI
- Reconciling Privacy with Provenance for Language Model Watermarks
- CoRTeC: A Differentially Private Release for Synthetic Tabular Data, Decodable by Any Generator
- A Bag of Tricks for Defending Deep Learning-based Face Recognition Pipelines against Backdoor Attacks
Robotics & embodied AI4
- Interactive World Simulation with Cosmos for Physical AINVIDIACode
- Open-HumanScale: An Open Recipe for Scaling Human and Robot Experience for Embodied Foundation ModelsNVIDIA
- MotoVerse: A Benchmark for Predictive Perception of Motorized Two-Wheelers in Dense Unstructured TrafficCode
- Jev with Foresight: Safe Language-Guided Autonomous Driving via a Language World Model