Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Researchers introduced FlexLAM, a latent action model that replaces fixed-capacity bottlenecks with variable-length codes trained via nested dropout, resolving the trade-off between discarding transition cues and preserv…
Researchers propose a physics-informed framework to discover anisotropic yield functions from full-field displacement and reaction force data, using a convex neural network representation. The method avoids direct stress…
Researchers propose a method to learn asynchronous denoising schedules for multi-representation diffusion models, improving visual synthesis. On ImageNet 256x256, their approach achieves FID 1.05 with 4x less training th…
Researchers propose a new framework integrating LLaVA and two-stage fine-tuning to improve composed image retrieval in fashion, addressing data scarcity and negative sampling limitations. The method enhances contrastive …
Researchers introduced cAPM, a continual AI-assisted pace-mapping system using active learning, which achieved 81% probability of localizing ventricular tachycardia targets within 5 mm accuracy using 4.5 pacing sites, co…
Researchers propose ProMUSE, a progressive multi-modal uncertainty-guided network for Alzheimer's disease classification that adaptively determines when additional imaging modalities are needed, reducing MRI/PET usage by…
Researchers developed a self-play reinforcement learning method for autonomous driving that uses only 30 minutes of human demonstrations, 2500 times less than imitation learning, to train policies that coordinate with hu…
Researchers derived a three-force decomposition of AdamW training dynamics explaining Weibull weight-scale parameter evolution in transformers, finding alignment force dominates growth. A spline displacement method recov…
Researchers at arXiv analyzed the performance of Med-DDPM, a 3D generative diffusion model for MRI synthesis, across three generations of NVIDIA GPUs. They identified inefficiencies in memory access and Tensor Core utili…
Researchers introduced the Integral Transform Network (ITNet), a unified architecture that subsumes convolution, attention, and recurrence as special cases of a learnable integral transform. ITNet matches or exceeds spec…
Researchers introduced MiqraBERT, a Sentence-BERT model finetuned from AlephBERT for detecting verse-level semantic similarity in Biblical Hebrew. The model improves distributional separation 2.7-fold over the pre-traine…
Researchers proposed LUCID, the first hallucination detection method for large language model-based knowledge graph reasoning, which jointly leverages LLM attention scores, KG semantics, and structural information. Exper…
Researchers introduced SSProNet, a secondary-structure-aware graph neural network for protein representation learning that uses energy-filtered hydrogen-bond graphs. The model outperforms existing graph-based methods on …
FocusOnFun team's Vortex system achieved 90.5% in the Preliminary Round and 'Excellent' overall performance with 'Outstanding' QA results at the Ho Chi Minh City AI Challenge 2025. The multimodal video retrieval system i…
A new human-in-the-loop pipeline measures how completely a computer science program covers international curricular guidelines, applied longitudinally to one accredited BSc program against CS2013 and CS2023. The program …
Researchers introduced self-function vectors to quantify aleatoric uncertainty in in-context learning for large language models, enabling more reliable prediction confidence and applications like hallucination detection.…
Researchers introduced LooseControlVideo, a framework that uses sparse, oriented 3D boxes as a proxy for spatial blocking in text-to-video generation, enabling intuitive control over layout and trajectory. Fine-tuned on …
Researchers introduced a decentralized coalition formation model driven by unilateral exit-and-join decisions, using the Aumann-Dreze value for local payoff calculations. The model links cooperative payoff allocation wit…
Researchers at arXiv propose a framework for selecting the most informative comparison pairs in preference-based post-training of large language models, showing that strategic pair selection can improve sample efficiency…
A new study measures the linearity of transformer feed-forward blocks, finding that linear recoverability (R²_lin) varies widely across blocks and is a learned property, not an architectural one. The analysis of GPT-2, P…