Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
Who Thinks First? Designing Productive Friction with Engage-to-Unlock GenAIAlphabet, Google DeepMind · arXiv (Cornell University)Oct 1LAST: Looped Audio Spectrogram TransformerGoogle DeepMind · arXiv (Cornell University)Oct 1Bridging Machine Learning and Algorithmic Information Theory, Part VII: Algorithmic Information Kernels and Kernel Discrepancies on Countable SpacesGoogle DeepMind · EntropyOct 1Polynomial-Time Cryptanalytic Extraction of Deep Neural Networks in the Hard-Label SettingGoogle DeepMind, Technology Innovation Institute · Journal of CryptologyOct 1Function-preserving watermarking of AI-generated proteinsGoogle DeepMind · Nature2 citationsSep 30Gumbel Straight Flow: Distilling Autoregressive Models into One-step Flow MapsGoogle DeepMind · arXiv (Cornell University)Sep 30Aligning Thoughts with Answers: Probability Rewards to Tame Thinking DriftGoogle DeepMind · arXiv (Cornell University)Sep 30Mathematical discovery and exploration can be done at scaleGoogle DeepMind · Proceedings of the National Academy of SciencesSep 30Conditional Generation of Creative Chess Puzzles with Diffusion ModelsGoogle DeepMind · arXiv (Cornell University)Sep 29TabFM-Auto: Self-Evolving Pipelines for Tabular Foundation ModelsAlphabet, Google DeepMind · arXiv (Cornell University)Sep 29Learning Dynamics of Continual Learning: A Unified View of Data Attribution, Forgetting, and Plasticity LossGoogle DeepMind · arXiv (Cornell University)Sep 27Generalizing equilibrium propagation to Lagrangian systems with arbitrary boundary conditions & equivalence with Hamiltonian echo learningGoogle DeepMind · Neuromorphic Computing and EngineeringSep 25STEVE: Stabilizing Textual Gradient-Based Prompt Optimization via Error-Driven Refinement and Regularized VerificationGoogle DeepMind · arXiv (Cornell University)Sep 20When Does Reasoning Help in Machine Translation? A Hierarchical Analysis of LRM Reasoning TracesGoogle DeepMind · arXiv (Cornell University)Sep 18Locating Hidden Failures Makes Long-Horizon Agents More ReliableAlphabet, Google DeepMind · arXiv (Cornell University)Sep 15Dream-RSI: Recursive Self-Improvement through Evolving WorldsGoogle DeepMind · arXiv (Cornell University)Sep 14From Connectivity to Rewards: Dense Reward Learning with Directed State GraphsGoogle DeepMind · arXiv (Cornell University)Sep 9Charts Are Beyond Pixels: Probing for Layer-Wise Chart Understanding and EditingGoogle DeepMind · arXiv (Cornell University)Sep 8Category-Native Solomonoff Approximation: From Algorithmic Geometry to Kernels, Operators, and InductionGoogle DeepMind · EntropySep 8Causal evidence that language models use confidence to drive behaviourGoogle DeepMind · Nature Machine IntelligenceSep 7Inducing Emergent Misalignment from Reward Hacks with Iterative DPOGoogle DeepMind · arXiv (Cornell University)Sep 6Mind the Gap: Exposing LLM Translation Blind Spots Using the AlphaMWE Multilingual Parallel CorpusGoogle DeepMind · arXiv (Cornell University)Sep 6Relaxed Activation Analysis of Dataflow Networks: A Clock Calculus for Machine Learning and Real-Time SchedulingGoogle DeepMind · Forum on Specification & Design Languages (FDL)Sep 1Learning Quantitative Automata Modulo TheoriesAlphabet, Google DeepMind · International Joint Conference on Artificial IntelligenceSep 1Visual General Intelligence: A White PaperGoogle DeepMind · arXiv (Cornell University)Aug 26