Researchers developed a Dynamic Neural Graph Encoder (DNG-Encoder) and the INR2JLS model, which uses dynamic graphs to boost neural network classification accuracy by approximately 10% on the CIFAR-10…
A new AI model developed using data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) aims to detect Alzheimer's disease early by handling class imbalance and refining feature selection. The…
Researchers developed a Sample-Adaptive Hyperbolic Graph Neural Network (SA-HGNN) that uses hyperbolic geometry to improve EEG-based depression detection by better capturing the hierarchical structure…
A new scheduler for large language model serving allows decode nodes to assist prefill phases, cutting P95 time-to-first-token by up to 81% and improving service-level objective attainment by up to 79…
Researchers introduced Conditional Co-Ablation (CoAx), a new method for AI interpretability that uncovers hidden circuit interactions by measuring how the impact of removing one component grows when a…
Researchers developed a communication framework using VQ-VAE that reduces latency by up to 79-fold with minimal accuracy loss, enabling efficient data transmission under spectrum constraints.…
Researchers introduced kNNGuard, a training-free guardrail for large language models that uses activation space to classify prompts, achieving up to 10x faster inference than fine-tuned methods. The s…
Researchers introduced MolSight, a graph-aware vision-language model that integrates topology and visual alignment to improve molecular image understanding, outperforming existing models in chemical v…
Researchers developed a semantic-aligned brain network framework that integrates semantics from large language models into brain disease diagnosis, improving stability and interpretability. The method…
Researchers introduced Rank-Then-Act (RTA), a framework that trains AI using video-derived ordinal signals instead of traditional rewards, achieving state-of-the-art results on benchmarks like PyBoy a…
Researchers introduced Decomposer, a framework that converts symbolic MIDI music into editable Strudel programs using synthetic data and reinforcement learning. The system outperforms closed-source la…
Researchers have developed a memory-efficient training stack for Mixture-of-Experts models that achieves 4.7x to 8.2x higher per-GPU throughput compared to a finely-tuned FSDP2 baseline, using fewer t…
Researchers used a consumer-grade EEG device to monitor cognitive load in online learning, achieving 78.5% accuracy in identifying difficult video segments with a machine learning model. However, the …
Researchers proposed a safety-constrained AI framework integrating perception and control to improve autonomous UAV navigation in dense environments. The framework uses a Lagrangian-based safe Proxima…
A new framework combining physics-informed neural networks (PINN) and graph neural networks (GNN) is transforming radio frequency (RF) map construction, offering superior accuracy and generalization f…
Researchers introduced a new metric, Composite Reward Observability Fraction (CROF), that improves AI model checkpoint selection in simulated environments like LunarLander. CROF combines Reward Observ…
COMFYCLAW, a new framework for workflow-based image generation, introduces skill evolution that allows AI agents to learn from past processes and improve over time. The system uses typed graph editing…
SOLiD, a framework for detecting deceptive AI responses, reduces undetected deception to 14% in models with 405 billion parameters, down from 34% in 1-billion-parameter models, while maintaining a 99%…
Researchers introduced PPRO (Profile-guided Personalized Retrieval Optimization), a framework that enhances memory retrieval for long-term conversational agents by integrating user profiles. The syste…
Researchers introduced the Geometry-aware R-Structured Kolmogorov-Arnold Network (GRS-KAN), a hybrid neural architecture that integrates geometric constraints via R-functions into the Kolmogorov-Arnol…