Researchers have introduced the Causal Discovery Foundation Model (CDFM), a unified framework for zero-shot structural causal inference that consistently outperforms traditional algorithms across dive…
LaunchSafe's Mako, an autonomous web exploitation engine built on the Self-Evolving Agentic Operating System (SE-AOS), achieved full-suite coverage on the XBOW validation benchmarks by exploiting 104 …
RepTran, a new AI repair method targeting Transformer feed-forward networks, achieved a 74.7% average repair rate across 18 fault benchmarks from CIFAR-100 and Tiny-ImageNet, outperforming random sele…
A University of Lausanne study found that an explainable machine learning pipeline assessing mental health risks in students produced stable results that reflected data construction rather than actual…
AMID, a new multi-agent framework for automating medical imaging model development, outperformed existing general-purpose machine learning engineering systems across 20 medical imaging challenge tasks…
Researchers have identified that the real cause of training instability in reinforcement learning with flow-matching policies is not iterative action generation but the vanilla sampling strategy, and …
Researchers have developed a multi-agent system using an expert-editor stepwise questioning strategy that significantly outperforms traditional LLM methods on long-document summarization, as demonstra…
A study testing 15 foundation models for mammography across 12 out-of-distribution datasets found that models like Mammo-FM and MaMA showed promise but struggled with generalization, while DINOv3, a v…
Researchers have developed HUMA (Hybrid Understanding for Multi-modal social Navigation), a new architecture that combines reinforcement learning with vision-language models to improve robot navigatio…
A new framework called the Distributed Agent System (DAS) introduces a device-edge-cloud architecture for fault-tolerant collaboration among AI agents, addressing reliability challenges in unpredictab…
Anamnesis, an open-source AI platform built on the Anthology and Alterity frameworks, allows researchers to prototype surveys on simulated virtual populations, generating immediate AI responses that m…
Researchers developed Heterogeneous Frame Dispatch Scheduling (H-FraDS), a method that optimizes NVIDIA edge GPUs for autonomous systems by achieving a 2.36x speedup over standard DLA execution, reach…
Researchers propose ARMOR (Anchor Rollout and Mixed Optimization for RL) to address reinforcement learning's over-optimization problem, which causes models to exploit training shortcuts and sacrifice …
Robo-ValueRL, a new reinforcement learning framework focused on value estimation, achieved an 86% success rate in millimeter-level precision tasks and 84% in generalized tasks after 240 hours of offli…
A new actor-critic reinforcement learning algorithm for deciding when to fine-tune AI models on resource-limited devices surpasses traditional fine-tuning approaches by over 4% in accuracy, achieving …
DiffUE introduces a new image protection technique that modifies semantic features rather than pixels, offering a defense against AI exploitation while preserving image quality. Tests on CIFAR-10 and …
Artificial intelligence is transforming electron microscopy into a dynamic, data-driven platform for nanoparticle analysis, enabling faster particle detection, segmentation, and morphology quantificat…
Self-supervised learning is pushing robots to learn without labeled data, but a new analysis reveals that evaluation protocols for continual self-supervised learning (CSSL) are inconsistent, making it…
Researchers have introduced Structured Thoughts, a new AI framework that organizes reasoning into two blocks to improve efficiency, achieving up to 8.08% performance gains in large language models. Th…
A novel framework redefines AI visibility metrics by introducing rank stability and structural sufficiency criteria, challenging traditional fixed-budget data collection methods. Applied to 30 platfor…