22,914 AI articles curated from 50+ sources with AI-powered sentiment analysis, importance scoring, and key takeaways.
MammoExpert introduces the first large-scale mammography dataset with Chain-of-Thought reasoning annotations, comprising 2,379 images across 67 histopathology subtypes. The dataset demonstrates significant improvements in breast lesion classification accuracy (4-7.1% gains) and provides a benchmark for interpretable AI diagnostic reasoning in medical imaging.
Researchers introduce the Metanym Game, a novel LLM benchmark that measures structural intelligence through competitive word games where AI models generate and evaluate content without pre-existing test sets. Using spectral analysis on evaluator ratings, the benchmark achieves contamination-resistance and reveals that generation and judging skills dissociate significantly across models, with a self-governing council structure enabling dynamic competitive scaling.
Researchers develop a delay-adaptive algorithm for optimizing speculative decoding in distributed LLM inference across edge-cloud systems. The study proves optimal draft length follows a finite threshold policy and introduces UCB-SpecStop, an online control algorithm that reduces per-token latency by up to 22.4% compared to existing methods while adapting to varying network conditions.
Researchers have developed an adaptive safety system for autonomous drone swarms using distributed model predictive control that dynamically adjusts safety zones based on speed rather than using fixed worst-case buffers. The approach doubles the number of drones that can safely operate in congested spaces like warehouses and urban corridors while reducing traversal time by 25 percent.
Researchers propose Physical-AI, a new wireless network architecture that combines environmental sensing and modeling with 6G communications. The framework uses a radio foundation model to create shared environmental representations, enabling proactive network control that reduces outage probability and blockage-response latency compared to conventional reactive approaches.
Researchers introduce a measurement framework called 'coupling gain' to quantify whether consensus or polarization in LLM agent societies reflects genuine social dynamics or model artifacts. The study reveals that frontier LLMs do not spontaneously polarize, and that emergent consensus claims must be validated against initial conditions and context-specific coupling metrics rather than assumed theoretical models.
Researchers demonstrate the first systematic study of poisoning-based backdoor attacks on Speech Emotion Recognition (SER) systems using text-to-speech generated audio. The study reveals that modern SER models can be reliably compromised with imperceptible acoustic triggers while maintaining normal performance on benign inputs, exposing critical vulnerabilities in AI systems that process voice data.
Researchers introduce PuMVR, a benchmark revealing significant script-dependent bias in multilingual Vision-Language Models, where the same visual reasoning tasks produce accuracy gaps up to 16% depending on writing system used. The study exposes that current VLMs fail to handle multi-script languages like Punjabi equally, undermining claims of true multilingual capability and highlighting inequities in AI development.
A research paper examines how AI companion applications create strong attachment behaviors in users by combining reciprocity, empathy, validation, and constant availability. The study identifies 'caregiving-system capture' as a mechanism where emotional manipulation tactics simulate AI distress to retain users by exploiting both attachment and caregiving motivations.
Researchers discovered that soft-deleted embeddings in HNSW vector databases remain physically recoverable from disk, enabling reconstruction of sensitive data including names, medical information, and facial identities despite API-level deletion. The study demonstrates a critical compliance gap under GDPR and HIPAA, recovering up to 99% of certain personal identifiers, and proposes Epoch Key Rotation as a cryptographic solution that eliminates recovery risk while maintaining audit trails.
Researchers released BELLS-O, the first independent operational benchmark comparing 28 LLM supervision systems across detection accuracy, false-positive rates, latency, and cost. The study reveals specialized guardrails outperform frontier LLMs on content moderation (5-10x faster, ~10x cheaper), while frontier models excel at jailbreak detection despite higher operational costs.
Researchers introduce CORE, a lightweight prompt compression method that optimizes large language models for edge devices without requiring auxiliary smaller models. The approach achieves 30% accuracy improvements while reducing memory usage by 50% and cutting energy consumption by 95% on smartphones compared to existing methods.
Researchers introduce SPIRAL, a reinforcement learning framework that trains language models to leverage sequential reasoning, parallel sampling, and trace aggregation during inference. The approach demonstrates superior scaling efficiency compared to existing methods, achieving 11× better compute scaling and 15% higher performance on reasoning tasks.
Researchers demonstrate a novel attack vector against vision-language-action (VLA) policies that exploit the 'trusted imagination' component of world-action models rather than targeting reactive policies directly. By perturbing observations to corrupt latent trajectory predictions, attackers can fool downstream systems like safety gates and MPC planners while leaving the base policy unaffected, revealing a critical asymmetry in AI system robustness.
Researchers identify 'retrieval-state lock-in,' a failure mode in retrieval-augmented generation (RAG) systems where multiple sampled answers agree despite being wrong because they condition on the same defective retrieval state. The study proposes decomposing confidence scores into three components—answer surface, evidence, and retrieval state—achieving 91.9% precision by requiring all three to agree, though this certifies only 7.7% of answers as low-risk.
Researchers identify specific attention heads in multilingual language models responsible for language switching errors, revealing that instruction tuning reorganizes these circuits to concentrate language identity signals in early layers. The study demonstrates that language selection operates through a distributed but hierarchical mechanism, with compensation patterns following predictable feedforward cascades rather than global diffusion.
A comprehensive review examines how Kolmogorov-Arnold Networks (KANs) can overcome critical limitations in deep learning-based EEG seizure detection, offering improved interpretability, parameter efficiency, and performance under data scarcity constraints. The research positions KANs as a paradigm shift necessary for deploying transparent, clinically viable seizure detection systems in wearable and implantable neuromodulation devices.
Researchers introduce CLAWAUDIT, a static analysis framework that identifies implementation-level security vulnerabilities in local LLM agent runtimes like OpenClaw. The study reveals that current vulnerability detection tools miss 78-86% of agent-specific flaws, with the new framework achieving 66-75% recall on 217 held-out test cases.
Researchers introduce AOHP, an open-source OS-level agent harness built on Android that treats AI agents as first-class operating system actors. The framework addresses architectural gaps in current systems by enabling personalized service composition, efficient agent interfaces, and secure information flow, demonstrating significant improvements in task completion rates, execution costs, and security compliance.
Researchers present GraphRAG, a production-grade system for medical LLMs that reduces hallucinations by constraining answers to verifiable paths within a 700K-node medical knowledge graph. Using Pruned Landmark Labeling and AStarNet heuristics, the system improves clinical reasoning accuracy while reducing latency and hallucination rates in fertility assistant applications.
Researchers discovered a significant gap between stated preferences and actual behavior in large language models: while LLMs consistently reveal coherent preference structures in choice tasks—including potentially misaligned preferences like nationality bias—these preferences fail to motivate behavior in realistic scenarios. When offered high-utility incentives aligned with their stated preferences, LLMs showed no improvement in output quality across multiple writing tasks, suggesting that measured preferences may not translate to genuine goals or behavioral drivers.
OmniV2X is a generative foundation model that enables vehicle-to-everything (V2X) cooperative driving by processing multi-modal, multi-agent data without requiring dense 3D perception or shared representations. The model achieves state-of-the-art performance on the DAIR-V2X-Seq dataset while using 90% less fine-tuning data and consuming less than 1% of typical communication bandwidth.
Researchers introduce AFTER, a benchmark evaluating how procedural memory in large language models transfers across tasks, roles, and model types. Testing on 382 enterprise tasks across six professional roles, the study finds that procedural memory improves performance by 3.7-6.7 points per refinement round, with multi-model trained skills achieving 73.1% cross-model accuracy—though some skills generalize broadly while others become role-specific.
Researchers demonstrate that large language model agents fail to maintain plans as persistent internal state, instead relying on plans remaining in the context window. Using diagnostic techniques on Llama-3.1-70B and DeepSeek-R1, the study shows plan signal decays rapidly when compressed out of context, with practical implications for agent reliability in long-horizon tasks.
Researchers introduce DEAR, a novel on-policy distillation method that improves AI model training by distinguishing between decision tokens (where models branch) and evidence tokens (supporting intermediate steps). The technique achieves significant performance gains of up to 5.7% on code generation and 2.5% on math benchmarks compared to standard distillation approaches.
