Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
A developer trained a small language model from scratch on an $8 ESP32-S3 microcontroller, demonstrating that training, not just inference, is possible on ultra-low-cost hardware. The project, inspired by the esp32-ai re…
Researchers at an undisclosed institution introduced a pipeline that uses a language model to automatically group features and MLP neurons into supernodes for circuit tracing, matching human annotator interpretability. O…
Researchers propose NOMADD, a post-hoc method to reduce concept drift in tabular models, applicable to trees, neural networks, and tabular foundation models. On the 18-dataset Drift-Resilient TabPFN benchmark, NOMADD imp…
Researchers introduced p-SpheretNorm, a one-parameter family of angular retractions for spherical residual streams in deep neural networks, showing that the exponential map used in Geodesic Normalization (GeoNorm) is jus…
Researchers propose Convex Hull Neighborhood Smooth Dual Generalization (CSDG), a new offline reinforcement learning method that explicitly separates in-sample value targets from local out-of-distribution corrections, ac…
Researchers propose CURV, a curriculum learning framework that improves chart question answering by reformulating it as multi-step visual grounded reasoning, achieving up to 20.50% improvements over baselines and up to 1…
A new arXiv preprint (2608.02829) finds that converting a pretrained 1.4B-parameter Pythia model into a 410M-parameter sibling works best through initialization transfer, not weight projection. The authors show that dens…
Researchers propose CHASE (Classification with Hierarchical Analysis and Structured Enforcement), a single-stage framework that uses a clinically driven three-level taxonomy of 9 anatomical regions, 17 sub-regions, and 2…
Researchers introduced DiffImaginE, a multimodal named entity recognition (MNER) verifier that formulates type verification as conditional latent diffusion inference, replacing deterministic imagine-and-compare methods. …
Researchers introduced the LLM Nominal Response Model (LLM-NRM), an option-aware psychometric framework that models the full distribution over answer choices to jointly estimate LLM ability and option-level item characte…
Researchers introduced Hi-TTRL, a test-time reinforcement learning framework that uses hints during sampling to regulate rollout consensus strength, improving reasoning in large language models without labeled data. The …
Researchers introduced Group-Relative Preference Backpropagation (GRPB), a method that converts same-task rankings into zero-sum advantages to train long-horizon video editing agents, and released the resulting 9B Crayot…
An AI-powered knowledge graph platform that maps SQL errors to conceptual gaps in database courses achieved 95.7% validity for extracted nodes and 63.8% fully correct triplets in an expert study, according to a paper rel…
Researchers propose OPTD (On-Policy Transition Distillation), a method that improves few-step diffusion language models by sampling from the student's own trajectories and using a frozen teacher to adaptively compress st…
BAP-SQL, a budget-aware observation planning method for agentic text-to-SQL, improves tight-budget success by 3.4/3.6 percentage points over matched SFT on the BIRD-derived setting while using 4.5/5.0% fewer tokens, acco…
Researchers introduced Local Credit Assignment (LoCA), a two-stage method for parameter-efficient fine-tuning of large language models that requires only one backward pass for calibration, after which tuning proceeds for…
Researchers introduced Self-Organising Digital Circuits, a new architecture using a topology-masked Transformer to configure Lookup Tables of Boolean gates, achieving near-perfect recovery (>99.99% accuracy) from soft er…
Researchers introduced FLARE (Few-shot Learning-based Adaptive Reflective Engine), a framework that uses reflective mechanisms and few-shot reference examples to optimize prompts for large language models, outperforming …
Researchers propose NANQ, a noise-aware mixed-precision non-uniform quantization framework for analog compute-in-memory (CIM), which models magnitude-dependent weight noise from measured eFlash CIM arrays to assign finer…
Researchers introduced OptR, an output-aware rotation method for INT2 KV-cache quantization that minimizes post-W_O attention-output error, improving long-context LLM inference efficiency. Across three models and five re…