Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
Automating Low-Risk Code Review at Meta: RADAR, Risk Calibration, and Review EfficiencyMeta · IEEE/ACM International Conference on Automated Software Engineering (ASE)Oct 7Customizing an LLM for Enterprise Software EngineeringAlphabet, Google DeepMind, Meta · IEEE/ACM International Conference on Automated Software Engineering (ASE)Oct 7REAP: Automatic Curation of Coding Agent Benchmarks from Interactive Production UsageMeta · IEEE/ACM International Conference on Automated Software Engineering (ASE)Oct 7Debug2Fix: Can Interactive Debugging Help Coding Agents Fix More Bugs?Microsoft · IEEE/ACM International Conference on Automated Software Engineering (ASE)Oct 7Synthesizing File-Level Data for Unit Test Generation with Chain-of-Thoughts via Self-DebuggingMicrosoft · IEEE/ACM International Conference on Automated Software Engineering (ASE)Oct 7SILENT: A New Lens on Statistics in Software Timing Side ChannelsTechnology Innovation Institute · Lecture notes in computer scienceOct 7ASPIRE: Agentic Safety & Prompt Injection Red-teaming EngineAlphabet · arXiv (Cornell University)Oct 6TRACE: Rollout-Guided Quantization-Aware Training for FP4 Reinforcement Learning of MoE Language ModelsAlibaba · arXiv (Cornell University)Oct 6CARE: Certifying Acceleration for Vision-Language-Action InferenceAlphabet · arXiv (Cornell University)Oct 6SpecGuard: Proving a Task Is Broken Before the Agent CheatsGoogle DeepMind · arXiv (Cornell University)Oct 6Matching Object or Relation? Tracing Abstract Reasoning Inside VLMsGoogle DeepMind · arXiv (Cornell University)Oct 6Personal-Agent Mediated Recommendation with Cross-Platform User HistoryMeta · arXiv (Cornell University)Oct 6Which and When to Admit: Gradient Admission for Data-Centric Small Language Model FinetuningMeta · arXiv (Cornell University)Oct 6Reinforcement Learning with Segment Reward Feedback under Linear Function ApproximationNVIDIA · arXiv (Cornell University)Oct 6Self-Referenced Social Preferences: Cooperation without Observing Others RewardsNVIDIA · arXiv (Cornell University)Oct 6Training Language Models To Be Coherent Decision-MakersMicrosoft · arXiv (Cornell University)Oct 6Continuous Memory MachinesSakana AI · arXiv (Cornell University)Oct 6Rethinking Cross-Tokenizer On-Policy Distillation: From Alignment Coverage to Supervision ReliabilityAlibaba · arXiv (Cornell University)Oct 6ASPIRE: Agentic Safety & Prompt Injection Red-teaming EngineAlphabet · arXiv (Cornell University)Oct 6DirectSpeech2LLM: A Simple End-to-End Framework to Mitigate Prompt Overfitting in Speech-LLMsMicrosoft · arXiv (Cornell University)Oct 6Reflections on the Security of In-Vehicle ConnectivityMeta · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications EngineeringOct 6Evaluating Safety Embedding Prefiltering for Analyzing Millions of LLM Agent Social Network Messages for Security and Safety HarmsNVIDIA · AIOct 6CCRA: Optimizing vision-language consistency via cross-layer regional attention alignmentTechnology Innovation Institute · Information FusionOct 6Load Testing for Machine Learning Model Serving Systems at ScaleMeta · Lecture notes in computer scienceOct 6MemCo: Memory-Centric Collaboration for Generalizing LLM Agents to Unseen EnvironmentsNVIDIA · arXiv (Cornell University)Oct 5