https://machinelearning.apple.com Apple machine learning teams are engaged in state of the art research in machine learning and artificial intelligence. Learn about the latest advancements. en Fri, 28 Aug 2026 00:00:00 GMT llms-not-consistently-bayesian https://machinelearning.apple.com/research/llms-not-consistently-bayesian Modern AI systems are being deployed in complex domains such as medicine, science, and law, where there is often not a single correct answer given the observed evidence. Such systems must be able to represent and update uncertain beliefs about the world as new evidence arrives to make rational decisions. We introduce the novel technique of studying LLMs as information processing rules and utilize the information processing gap—the deviation from Bayes updates—to study the internal (in)consistencies of how LLMs update their probabilistic beliefs from evidence. Our extensive experiments evaluate… Fri, 28 Aug 2026 00:00:00 GMT agent-seer-synthesizing-scenarios https://machinelearning.apple.com/research/agent-seer-synthesizing-scenarios Evaluating AI agents that use external tools requires realistic test scenarios that capture how practitioners compose tools and iterate across conversation turns. Constructing such scenarios by hand demands deep domain expertise, does not scale across tool ecosystems, and produces static benchmarks that cannot track evolving APIs. We observe that tool specifications—function names, natural-language descriptions, and typed parameter schemas—already encode sufficient semantic information to synthesize realistic evaluation scenarios without manual curation or live tool execution. Agent Seer… Fri, 28 Aug 2026 00:00:00 GMT rubric-based-alignment https://machinelearning.apple.com/research/rubric-based-alignment Designing effective reward signals for open-domain question answering is challenging because high-quality responses must simultaneously satisfy multiple aspects of answer quality that are difficult to capture with a holistic scalar objective. We introduce a rubric-based reward framework that generates query-specific rubrics grounded in retrieved evidence and decomposed into multiple quality dimensions, providing fine-grained supervision during post-training. Averaged across three evaluation axes (composition, grounding, and instruction-following), our approach improves over the… Thu, 27 Aug 2026 00:00:00 GMT proof-gen-optimized-distillation https://machinelearning.apple.com/research/proof-gen-optimized-distillation Supervised fine-tuning on teacher-generated trajectories is the standard first stage for distilling tool-calling capabilities into deployable models. Post-training pipelines that drive shipped tool-calling agents re-run this stage on a daily or weekly cadence, paying the frontier-teacher cost each cycle, yet the mechanism is generate-and-filter (keep the teacher’s passing trajectories, discard the rest) and each cycle leaves behind the same hard scenarios because failures supply no signal. On τ 2-bench, 57% of teacher trials fail, two-thirds of them near-misses (most tool calls correct, undone… Wed, 26 Aug 2026 00:00:00 GMT idea-prune-pipeline https://machinelearning.apple.com/research/idea-prune-pipeline Recent advancements in large language models have intensified the need for efficient and deployable models within limited inference budgets. Structured pruning pipelines have shown promise in token efficiency compared to training target-size models from scratch. In this paper, we advocate incorporating enlarged model pretraining, which is often ignored in previous works, into pruning. We study the enlarge-and-prune pipeline as an integrated system to address two critical questions: whether it is worth pretraining an enlarged model even when the model is never deployed, and how to optimize the… Wed, 26 Aug 2026 00:00:00 GMT relightable-gaussians-3d-generation https://machinelearning.apple.com/research/relightable-gaussians-3d-generation High-fidelity image-to-3D generation requires a 3D representation that captures both geometry and appearance. To support relighting and integration into standard rendering pipelines, the representation should include physically based rendering (PBR) modalities such as albedo, metallic-roughness, and surface normals. We propose Luce, a 3D representation that unifies geometry and PBR materials within a voxelized multimodal Gaussian cloud, using dedicated Gaussian primitives for each modality. A variational autoencoder compresses this representation into a unified material-aware latent space. A… Wed, 26 Aug 2026 00:00:00 GMT starflow2-multimodal-generation https://machinelearning.apple.com/research/starflow2-multimodal-generation Unified multimodal models that understand, reason over, and generate interleaved text–image sequences remain structurally fragmented: existing approaches either sacrifice visual fidelity through discrete tokenization, impose structural asymmetry by combining causal text generation with iterative diffusion-based denoising, or degrade pretrained understanding when adapting vision-language models for generation. We observe that autoregressive normalizing flows are autoregressive Transformers—sharing the same causal mask, KV-cache mechanism, and left-to-right structure as LLMs—making them the most… Tue, 25 Aug 2026 00:00:00 GMT internalized-visual-thinking https://machinelearning.apple.com/research/internalized-visual-thinking Multimodal large language models increasingly use visual chain-of-thought (Visual CoT) to reason about spatial, temporal, and embodied environments. By generating intermediate reasoning images, Visual CoT provides an intuitive mechanism for visual foresight but introduces substantial inference overhead, which is particularly problematic for proactive video reasoning. We ask whether models can learn to think visually during training while reasoning directly at inference. We introduce Internalized Visual Thinking (IVT), a post-training framework that jointly optimizes textual prediction and… Mon, 24 Aug 2026 00:00:00 GMT multilingual-knowledge-transfer-lexical-interventions https://machinelearning.apple.com/research/multilingual-knowledge-transfer-lexical-interventions Cross-lingual knowledge transfer is critical for building high-performing multilingual language models for languages with insufficient training data. When target language data is scarce, the knowledge required for many downstream tasks involving scientific reasoning, commonsense inference, and world knowledge must be acquired primarily from the high-resource language, making effective knowledge transfer essential. Existing methods for improving such cross-lingual knowledge transfer require large amounts of parallel data, translation systems, auxiliary models, or additional training stages that… Thu, 20 Aug 2026 00:00:00 GMT scaling-laws-mixture-pretraining https://machinelearning.apple.com/research/scaling-laws-mixture-pretraining As language models scale, the amount of data they require grows – yet many target data sources, such as low-resource languages or specialized domains, are inherently limited in size. A common strategy is to mix this scarce but valuable target data with abundant generic data, which presents a fundamental trade-off: too little target data in the mixture underexposes the model to the target domain, while too much target data repeats the same examples excessively, yielding diminishing returns and eventual overfitting. We study this trade-off across more than 2,000 language-model training runs… Thu, 20 Aug 2026 00:00:00 GMT