Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Photoroom's SDXL LoRA fine-tuning for a product photography model silently corrupted its adapter weights over six days due to a bf16 gradient accumulation issue. The custom training loop, forked from an internal repo two…
Utilora has built a fully client-side Retrieval-Augmented Generation (RAG) system called Personal RAG that runs entirely within a web browser, enabling private AI search over local files without any server, API, or cost.…
Researchers Haftu W. Fentaw and two coauthors introduced a Multi-task Soft Actor-Critic framework that simultaneously learns optimal pulse sequences and discovers problem-specific evolution time and control segment numbe…
Euimin Lee and Shiho Kim formulated step-wise measurement-shot allocation inside depth-1 Recursive QAOA as a sequential decision problem in arXiv paper 2605.26544. The authors evaluated a hand-crafted heuristic and a tab…
A new arXiv preprint proposes an end-to-end deep learning framework that directly learns portfolio weights by combining Long Short-Term Memory networks, Graph Attention Networks, and financial news sentiment analysis. Th…
Researchers Juan Cruz-Benito and Ismael Faro adapted Microsoft's QuantumKatas exercises from Q# to Qiskit, creating a benchmark of 350 tasks across 26 categories to evaluate large language models on quantum code generati…
Wealth management firms are adopting data-driven personalization at scale, moving beyond basic digital tools toward operating models shaped by data science, artificial intelligence, automation, and portfolio engineering.…
Researchers from the National University of Singapore, MIT CSAIL, A*STAR, and SMART have developed MEMO, a framework that trains a separate, small language model to internalize new knowledge from documents without modify…
A developer built a full-stack sentiment analysis web app as their first machine learning project, using a Random Forest model with Scikit-Learn and TF-IDF vectorization on the backend and React on the frontend. The app …
A team of Bayesian statisticians has developed a method to fit million-parameter models using serverless GPU computing for pennies, dramatically reducing the cost and complexity of Markov chain Monte Carlo (MCMC) inferen…
A benchmark comparing Chronos-Bolt and Toto zero-shot forecasting models on Prometheus and OpenSearch telemetry data found that both models outperformed naive baselines on periodic memory signals but struggled with heavy…
A developer with over a decade of AI experience has published a learning roadmap for complete beginners, warning that starting with generative AI tools like ChatGPT is the biggest mistake newcomers can make. The roadmap …
Microsoft Research's ARTIST framework trains LLM agents to discover when and how to call tools through reinforcement learning, using only outcome-based rewards rather than step-by-step annotations. At 7B scale, ARTIST ou…
FlashLib, a new GPU library for classical machine learning operators, achieves speedups of up to 208× over cuML on Hopper GPUs for algorithms including KMeans, KNN, and PCA. The library is designed to accelerate the "int…
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 …
MiniMax released the M2 series of Mixture-of-Experts language models, featuring 229.9 billion total parameters with only 9.8 billion activated per token, designed for agentic deployment. The models use agent-driven data …
Researchers have introduced GEM (Geometric Entropy Mixing), a framework that reformulates LLM data curation as a variational problem on a hypersphere to overcome flaws in human taxonomies and Euclidean clustering. By dec…
Researchers introduced SPEAR (Sandboxed Prompt Engineer with Active Roll-back), a code-augmented agentic optimizer that autonomously rewrites prompts using a Python sandbox for structural error analysis, outperforming ex…
A new study finds that vector representations exported from large language model summarization systems can leak sensitive information, even when source documents remain access-restricted. Researchers auditing clinical di…
Researchers have introduced Neural Bayesian Sequential Routing (NBSR), a framework that models neural inference as active evidence accumulation over a hierarchical Directed Acyclic Graph using a Dirichlet-Categorical con…