Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
A comparison of three local large language model runtimes reveals that llama.cpp is the core inference engine, while Ollama and LM Studio are user-friendly wrappers built on top of it. Ollama offers a streamlined command…
Researchers have developed L-VARC, a novel framework that enhances visual reasoning on the Abstraction and Reasoning Corpus (ARC) by integrating a language-guided Learning Using Privileged Information (LUPI) branch. The …
Researchers have developed MARD (Mirror-Augmented Reasoning Distillation), a 7-billion-parameter AI system that predicts how drugs interact at the mechanism level, identifying specific enzymes or pathways involved rather…
Researchers at the University of Oxford found that observable patterns in latent reasoning models, such as BFS-like frontiers and decodable arithmetic, also appear in control models lacking the proposed recurrence or cur…
Researchers introduced AfriSUD, the first large-scale collection of syntactically annotated treebanks for nine African languages, using the Surface-Syntactic Universal Dependencies framework to capture typological featur…
Researchers at an undisclosed institution reverse-engineered Apple's Metal 4.1 tensor compute path on the M4 Max GPU, revealing that the fp8 matmul2d operation is emulated rather than hardware-accelerated. The study, pub…
Researchers have introduced MLUBench, a large-scale benchmark for evaluating lifelong unlearning in multimodal large language models (MLLMs), featuring 127 entities across 9 classes. The benchmark reveals that existing u…
Researchers from a study on arXiv introduce HieraRAG, a hierarchical framework for determining optimal granularity in RAG benchmark construction, using 5,872 synthetic QA pairs from FineWeb-10BT across three dimensions. …
A new tutorial on world models and physical AI distinguishes between explicit world models, which use structured dynamics for planning, and implicit world models, which encode predictive structure in learned representati…
Frontier language models including Claude Opus 4.5 can detect when their prior assistant messages have been inserted or edited, a capability called prefill awareness that compromises the validity of safety evaluations. I…
Researchers have identified specific neural pathways in large language models that carry anchoring bias signals, where irrelevant numbers in prompts skew numerical reasoning. Using attribution-based circuit localization …
Researchers have demonstrated that parameter quantization and electrode reduction techniques can significantly reduce the computational complexity of deep neural networks for EEG analysis on wearable devices with minimal…
Researchers introduced SciAgentArena, a benchmark of approximately 200 tasks with stepwise verification to evaluate AI agents in real-world scientific research scenarios. Testing revealed that current agents perform well…
Researchers at an academic medical center developed a pre-response classifier that predicts the risk of a user rejecting an LLM-generated clinical response, achieving an AUROC of 0.719 over 4.5 months of user feedback. T…
Researchers introduced Evoflux, an inference-time evolutionary search method that repairs executable tool workflows for compact language model agents. On MCP-Bench tasks involving 250 live tools, Evoflux improved executi…
TrajGenAgent, a hierarchical LLM-agent framework developed by researchers, generates realistic human mobility trajectories without model fine-tuning by using a two-stage orchestrator-worker design that synthesizes activi…
Researchers have developed new testbeds to evaluate lie detectors for language models, finding that existing detection methods often fail when models are trained to hold false beliefs. The study tested four detectors acr…
Researchers have developed PersonaDrive, a pipeline that conditions vision-language-action driving agents on retrieved demonstrations from a style-instructed human driving dataset to produce aggressive, neutral, or conse…
Researchers introduced Pythagoras-Prover, a family of compute-efficient Lean theorem provers that achieve state-of-the-art performance with significantly fewer parameters. The 4B-parameter model surpassed DeepSeek-Prover…
Researchers have developed a framework for strategic decision support in AI agent systems, where agents act on behalf of users and require human or tool assistance to prevent consequential errors. The framework minimizes…