OpenAI is in early discussions to give the US government a 5% stake, CEO Sam Altman said, aiming to align AI benefits with national interest and improve relations with the Trump administration.…
Procedural Memory Distillation (PMD) advances reinforcement learning by transforming experiences across episodes into actionable intelligence, outperforming previous models like SDPO by 3.8-5.5% on SC…
New research quantifies how quantization affects AI model decision boundaries, showing that 8-bit weight quantization preserves all test labels on the digits benchmark with a boundary-mask Jaccard of …
A new meta-benchmarking framework evaluates AI models specifically for financial domains, organizing 452 benchmarks into 38 banking domains and using a weighted scoring system to provide more relevant…
CausalSTeward (CAST), a multi-agent AI framework that integrates human expertise with machine intelligence, aims to improve causal model learning from high-dimensional data by partitioning variables i…
Researchers have developed a model-agnostic framework that integrates privacy-enhancing coded computing with defense mechanisms to protect distributed machine learning systems from privacy breaches an…
Researchers have identified flaws in the common practice of using random dataset splits for AI evaluation, which fails in fields like aerial surveillance and agriculture due to spatial and temporal da…
Large language models (LLMs) show promise in automating cloud access control policy generation, with reasoning-enabled models achieving 93.7% accuracy versus 45.8% for standard models, but full autono…
Conditional Inference Forests (CIF) rank 4th among 17 classification methods and 3rd among 18 regression methods across multiple datasets, offering unbiased feature ranking despite runtime costs that …
Lynx, a new system for large language model inference, reduces latency by splitting the KV cache into Anchor and Residual streams, enabling speculative decoding. It improves time-to-first-token by up …
Researchers used deep learning to virtually expand a 4-microphone tetrahedral array into a 32-microphone spherical array, achieving a root mean square error of 0.432 on the STARSS23 dataset. The metho…
Researchers introduced Prompt Coverage Adequacy, a new software testing criterion that evaluates how well prompts guide development by leveraging LLMs' attention mechanisms, detecting over 30% more fa…
CLAP (Closed-Loop Agent Post-training) offers a structured method for post-training AI agents, but trials show modest gains and mixed results across manufacturing scenarios, with risks including high …
New research proposes a Bayesian deep learning framework called MP-TTBDL that uses a residual recurrent gated unit to model hardware impairments in massive MIMO receivers, achieving lower channel esti…
Choosing between AI-specific majors and traditional STEM degrees shapes careers in artificial intelligence. A hybrid approach combining AI specialization with a STEM foundation offers adaptability as …
Researchers introduced set diffusion, a new class of language models that generate flexible-position, flexible-length token sets instead of fixed-length sequences, achieving faster inference and bette…
Researchers introduced an attention-based diffusion model (ADMC) for multimodal emotion recognition that handles missing data by independently training networks per modality, achieving state-of-the-ar…
Researchers introduced DecompRL, a reinforcement learning algorithm that breaks complex problems into smaller sub-tasks for large language models, reducing GPU token costs by approximately 50 times. T…
Researchers introduced Adaptive Reparameterized Time (ART), a method that dynamically adjusts diffusion sampling timesteps to improve sample quality without overhauling existing systems. ART uses rein…
ContraFix, a new automated vulnerability repair system using GPT-5-mini, achieves a 92% resolution rate on SEC-Bench and 73.8% on PatchEval across Go, Python, and JavaScript. Its contrastive runtime a…