Full-text search across 7046 articles. Combine with topic and date filters; results sorted by relevance.
Prima.cpp, a new software tool, enables large language models with up to 70 billion parameters to run efficiently on ordinary home computers using pipelined-ring parallelism, achieving 5 to 17 times lower token processin…
The SEMQ Group, led by CEO Andrés Mac Allister, introduced Symbolic Embedding Multi-Quantization (SEMQ), a new method for AI model efficiency that separates semantics from numeric representation. Initial tests on the Ban…
Researchers have demonstrated that forcing neural networks to adopt low-dimensional representations significantly enhances their ability to generalize, according to a new study. Using information-theoretic measures, the …
A breakthrough in AI concept representation uses Laguerre Geometry to precisely map concepts as regions rather than points, reshaping interpretability. Researchers have developed Geometric Lens, a method that offers unpr…
New research shows depth is more important than width in neural networks for approximating analytic functions, challenging traditional thinking. The study derived new approximation rates expressed as O(N^{-CL^τ}), where …
Google has unveiled the Magic Pointer, a feature powered by its Gemini AI that reads screens in real-time and suggests actions, aiming to redefine user interaction with Googlebook laptops. The system uses natural languag…
Researchers have found that the multimodal language model LLaVA 1.6 can replicate six of seven categories of naming errors seen in stroke-induced aphasia, matching real-world clinical data in 97.8% of configurations, tho…
NSNQuant, a calibration-free vector quantization method developed by researchers, promises up to 3x throughput gains for language model inference by compressing the key-value cache without relying on calibration datasets…
Git-Assistant, a new tool combining large language models with automated planning, aims to simplify complex git operations by translating natural language commands into precise actions. A systematic evaluation shows the …
A new approach using sparse autoencoders (SAEs) is revolutionizing out-of-distribution (OOD) detection by extracting interpretable features from intermediate neural network layers, achieving record performance on OOD det…
SPARC-Net, a novel neural network architecture, claims to overcome the limitations of Physics-Informed Neural Networks (PINNs) in solving stiff and shock-dominated partial differential equations, achieving error reductio…
A new framework grounded in generalized extrinsic information transfer (GEXIT) functions redefines diffusion model training by defining cross-entropy as an integral of local information-theoretic derivatives along the no…
A study of 40 million papers found that scientists using AI publish more and advance faster, but AI-heavy research clusters around the same problems, reducing scientific diversity. The findings, published in Nature, high…
AI model performance increasingly depends on data quality rather than compute power, as researchers shift focus from scaling model size to refining datasets. This trend affects real-world applications like autonomous veh…
Researchers introduced ReRanking Preference Optimization (RRPO), a method that uses reinforcement learning from LLM feedback to improve reranker relevance for AI-generated answers, outperforming strong baselines like Ran…
Researchers developed a two-stage AI pipeline that estimates wind conditions and uses reinforcement learning to improve drone trajectory tracking by 48% in turbulent winds. The method, tested in simulations, outperforms …
Researchers introduced two CUTLASS-based kernels that accelerate the SwiGLU activation function in large language models by up to 2.47x on NVIDIA H100 GPUs, shifting workloads from memory-bound to compute-bound and achie…
Researchers introduced K9-Bench, a benchmark evaluating AI's ability to understand canine behavior through 5,000 question-answer pairs across 907 videos. Current AI models struggle with subtle animal interactions, highli…
Researchers introduced a Gibbs algorithm that simplifies posterior sampling in Bayesian inversion, offering reliable convergence guarantees and computational efficiency. The method, verified through numerical simulations…
Cloudflare announced that key members of Ensemble AI are joining the company to accelerate AI infrastructure development. Ensemble AI, founded in 2023, focuses on making large AI models faster and more cost-effective thr…