Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Researchers introduced LongAV-Compass, a benchmark for evaluating minute-long audio-visual generation across text, image, and video conditioning modalities. The benchmark contains 284 test cases and assesses over 20 fine…
Researchers have developed RadarSim, a unified differentiable renderer that uses multimodal neural fields to generate high-resolution Doppler radar range images by leveraging camera data. The system addresses the low spa…
Researchers have developed BEAP, a black-box adversarial prompting attack that exploits vulnerabilities in text-to-image diffusion models that have undergone machine unlearning. The attack uses a large language model to …
Researchers have developed CE-FedGNN, a federated graph neural network framework that enables organizations to collaboratively learn from distributed graph data without sharing raw data or frequent embedding exchanges. T…
A new study benchmarked twelve deep learning architectures across four model families—convolutional neural networks, vision transformers, hybrid CNN-transformer backbones, and vision-language models—for multi-disease ret…
Researchers have developed DecoupleGen, a method that uses personalized text-to-image diffusion models to generate training images of objects in uncommon contexts, such as a beach ball on a road. The approach aims to cor…
A new survey reframes alignment tuning for large language models as a data pipeline design problem, decomposing data construction into three stages: response synthesis, preference evaluation, and preference instantiation…
A user has published a guide for running llama.cpp with Vulkan on a 2013 Mac Pro 6,1 equipped with dual FirePro D700 GPUs running Ubuntu. The guide explains that the machine's 12 GB of aggregate VRAM must be treated as t…
Researchers introduced FML-Bench, a benchmark of 18 machine learning research tasks across 10 domains, to isolate the impact of agent strategy from execution infrastructure on AI research agent performance. Testing six a…
A developer has published a 40-line Python script that logs machine-learning option-pricing forecasts from the Helium MCP REST API to a CSV file, enabling Brier-score calibration after contracts expire. The open-source r…
A developer built Micro-Expert-Router, a Rust inference engine that streams Mixture-of-Experts model weights directly from NVMe SSDs using io_uring with O_DIRECT, eliminating the need for GPU VRAM. The engine targets mod…
NVIDIA's C++ template-based CUTLASS library for GPU kernels suffers from compile times of up to 20 seconds for a single kernel and over 17 minutes for full builds, prompting a shift toward Python-embedded DSLs. A tracing…
Two engineers in Dubai built an AI memory engine inspired by the hippocampus that reduced retrieval-augmented generation (RAG) costs by 10x. The system uses sparse binary codes—40 active bits out of 8,192 slots—to achiev…
Researchers at the Karlsruhe Institute of Technology (KIT) in Germany have demonstrated that standard WiFi networks can identify individuals with 99.5% accuracy by analyzing unencrypted beamforming feedback data using ma…
Sparse embeddings represent text by assigning binary values based on token presence in a vocabulary dictionary, using methods like TF-IDF and BM25 for keyword-based retrieval. While effective for exact text matching, spa…
Alibaba Cloud has joined the PyTorch Foundation as a Platinum member, gaining a seat on the foundation's Governing Board and a position on its Technical Advisory Committee. The Chinese cloud computing giant, which develo…
HASH researchers argue that building a process foundation model (PFM) requires a novel architecture, Dynamic Graph-JEPA, which extends Graph-JEPA to handle partially-observed, typed, temporal graph state. They claim exis…
Chan Zuckerberg Biohub released ESMC, ESMFold2, and ESM Atlas under an MIT license this morning, and Sheaf v0.11 shipped support for both model backends within twelve hours. The same-day integration was possible because …
Open-weight AI models are now 3–6 months behind frontier cloud models in benchmark performance, but the gap is closing fast enough that local AI has become a viable infrastructure decision for cost, privacy, and agentic …
Ollama and LM Studio now allow users to run open-weight AI models like Qwen 3, Gemma 3, and DeepSeek-R1 locally on consumer hardware without API keys or monthly fees. The tools support quantized models in GGUF format, en…