Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
A controlled comparison of byte-pair encoding (BPE) and Unigram-LM tokenizers on chemistry SMILES reveals that the two algorithms build near-disjoint subword vocabularies across all 22 tested conditions, with cross-algor…
Researchers evaluated UAV detection and tracking under synthetic fog, finding that fog degrades performance primarily through missed detections. Fog-inclusive training improved robustness most consistently, while test-ti…
Researchers propose TurnOPD, a turn-level budgeting strategy for efficient on-policy distillation of long-horizon language agents. The method addresses inefficiencies in vanilla agent OPD by using adaptive rollout-depth …
Researchers derived smooth Monte-Carlo estimators for Minkowski functionals of neural fields, achieving 1-3% accuracy in 2D and 3D topology estimation at 250x speedup over persistent homology. However, in 3D neural-SDF f…
Researchers introduced ArtisanCAD, an industrial-level CAD agent that uses expert-grounded knowledge distillation to convert ambiguous text prompts into executable CAD programs. The system leverages a CAD intermediate re…
Researchers introduced SPANUQ, a lightweight probe for span-level uncertainty quantification in large language models, achieving 10-20x faster inference than sampling-based methods while outperforming them in uncertainty…
Researchers at arXiv introduced HARVEY, a defense against neural backdoor attacks that learns a reference model for poisonous samples instead of benign ones. By identifying poisonous samples more accurately than prior me…
A new arXiv paper demonstrates that standard importance weighting fails to correct for selection bias in machine learning when selection depends on unobservables correlated with outcomes. The authors adapt Heckman's 1979…
Researchers introduced Light-Omni, a multimodal agent framework for reflexive video understanding that uses dual contextual states to avoid costly iterative reasoning. It achieved a 2.4% accuracy gain over M3-Agent with …
Researchers propose MuCoDi, a pretraining framework that distills knowledge from multiple large pathology foundation models into compact edge-oriented encoders, achieving near-teacher performance with orders-of-magnitude…
Researchers introduced FASR++, a diffusion-model-based super-resolution algorithm that leverages multiple low-quality images from surveillance cameras to enhance facial features and improve face recognition. The method a…
Researchers introduced FedPPO-PG, a federated multi-agent reinforcement learning framework for transient stability control in smart grids. The system achieved 100% stabilization in simulations on the IEEE 39-bus benchmar…
Researchers introduced InvWeaver, a neuro-symbolic framework that synthesizes loop invariants for programs with multiple interacting loops by exposing inter-loop dependencies and propagating proof obligations. The system…
Researchers introduced REVIVE, a multi-modal framework for detecting and recovering from vandalism-induced occlusion attacks on autonomous vehicle cameras. The system combines detection, segmentation, and type-aware inpa…
Researchers developed a vision-language framework using Low-Rank Adaptation (LoRA) on a frozen BiomedCLIP backbone to detect severe adverse events in wound monitoring. The system integrates clinical notes and wound descr…
Researchers introduced Algorithmic Core Extraction (ACE) to isolate compact subspaces, or algorithmic cores, that are necessary and sufficient for a task and recur across independently trained transformers. Across synthe…
Researchers introduced Self-Review Reinforcement Learning (SRRL), a training framework that embeds a self-review step into each RL episode to help large language models learn from sparse feedback. SRRL outperformed stand…
Researchers proposed a method to estimate observation parameters and noise in inverse problems using a diffusion-based prior model within a Bayesian framework. The approach enables optimal estimation of both parameters a…
Researchers show that incorrect L0 hyperparameter settings in Sparse Autoencoders (SAEs) cause feature mixing and degenerate solutions, preventing proper disentanglement of LLM internal features. They propose a proxy met…
Researchers propose IMR, a novel method for multi-agent trajectory prediction that combines mode-world weighted regression loss and an iterative decoder to improve mode diversity and prediction accuracy. The method ranks…