Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
Hair-Trigger Alignment: Black-Box Evaluation Cannot Guarantee Post-Update AlignmentAlphabet, Google DeepMind · arXiv (Cornell University)Jan 29DP-λCGD: Efficient Noise Correlation for Differentially Private Model TrainingAlphabet · arXiv (Cornell University)Jan 29Hair-Trigger Alignment: Black-Box Evaluation Cannot Guarantee Post-Update AlignmentAlphabet, Google DeepMind · arXiv (Cornell University)Jan 29Visual-Guided Key-Token Regularization for Multimodal Large Language Model UnlearningAlphabet · arXiv (Cornell University)Jan 29Adaptively Robust Resettable StreamingAlphabet · arXiv (Cornell University)Jan 29Noisy Pairwise-Comparison Random Search for Smooth Nonconvex OptimizationGoogle DeepMind · arXiv (Cornell University)Jan 29ECO: Quantized Training without Full-Precision Master WeightsAlphabet · arXiv (Cornell University)Jan 29A benchmark of expert-level academic questions to assess AI capabilitiesAlphabet, Anthropic, Cohere, Google DeepMind, Microsoft, OpenAI · Nature63 citationsJan 28PhaseCoder: Microphone Geometry-Agnostic Spatial Audio Understanding for Multimodal LLMsAlphabet · arXiv (Cornell University)Jan 28Dissecting Multimodal In-Context Learning: Modality Asymmetries and Circuit Dynamics in modern TransformersAlphabet · arXiv (Cornell University)Jan 28RIFS: Run-Time Invariant Function SpecializationAlphabet · ACM SIGPLAN International Conference on Compiler Construction (CC)Jan 28Does AI Now Represent a Paradigm Shift?Alphabet · Communications of the ACMJan 28Predicting Adaptively Chosen Observables in Quantum SystemsAlphabet · PRX QuantumJan 28Interdisciplinary expertise to advance human-centered explainable AIAlphabet · Social Science & Medicine1 citationJan 27Dynamic Cogeneration of Bug Reproduction Test in Agentic Program RepairAlphabet · arXiv (Cornell University)Jan 27Thought-Transfer: Indirect Targeted Poisoning Attacks on Chain-of-Thought Reasoning ModelsAnthropic, Google DeepMind, OpenAI · arXiv (Cornell University)Jan 27Regularized $f$-Divergence Kernel TestsAlphabet · arXiv (Cornell University)Jan 27Dynamic Cogeneration of Bug Reproduction Test in Agentic Program RepairAlphabet · arXiv (Cornell University)Jan 27SAGE: Steerable Agentic Data Generation for Deep Search with Execution FeedbackAlphabet · arXiv (Cornell University)Jan 26Practical block encodings of matrix polynomials that can also be trivially controlledAlphabet · arXiv (Cornell University)Jan 26SAGE: Steerable Agentic Data Generation for Deep Search with Execution FeedbackAlphabet · arXiv (Cornell University)Jan 26iPDB -- Optimizing Semantic SQL QueriesAlphabet · arXiv (Cornell University)1 citationJan 23Hardness of Learning Boolean Functions from Label ProportionsAlphabet · ACM Transactions on Computation TheoryJan 22Predictive Analytics for Data Pipeline Reliability: Machine Learning Models to Forecast ETL Failures and BottlenecksAlphabet · International Conference on Communication, Computing and Electronics Systems (ICCCES)Jan 21Hybrid Convolutional Neural Network Model for Robust and Sustainable Underwater Image EnhancementAlphabet · International Conference on Communication, Computing and Electronics Systems (ICCCES)Jan 21