summarizeddata
topic
a new method called goen uses multi-scale features and mahalanobis distance to beat deep ensembles on out-of-distribution detection, while revealing that centerloss harms performance.
nvidia releases diffusion language models that generate multiple tokens in parallel and refine them, offering faster inference and flexible modes.
the pytorch foundation seeks new ambassadors to lead local communities, create tutorials, and mentor users, with applications open until june 18, 2026.
a 3b specialized model outperformed frontier apis at lower cost, showing training alignment matters more than parameter count.
a small fraction of autoregressive model requests can loop into repetition, inflating batch inference time by over 40% and exposing a blind spot in standard benchmarks.
a tutorial shows how to give gemma 4 tools for local filesystem exploration and restricted python execution, making it more agentic.
a new method trains language models to predict which research ideas will work better without running experiments, using a dataset of 11,488 idea pairs from paperswithcode.
a new proof shows that feature rankings cannot be simultaneously faithful, stable, and complete when features are correlated, and proposes an ensemble method as a solution.
a curated look at top small language models under 7b parameters on hugging face, with benchmarks and code to get started.
a neural estimator learns pairwise conditional mutual information from pretrained masked diffusion models, enabling faster parallel decoding by identifying independent variable subsets.
a new transformer variant fixes training instability in looped models by distributing signals across layers and reusing attention states, enabling deeper iteration without extra parameters.
a curated list of github repositories covering strategies, frameworks, and tools for systematic quantitative trading development.
practical patterns for running totals, sessionization, cohort analysis, and percentile ranking using sql window functions.
learn to use the mimesis python library to replace sensitive personal data with realistic fake data for safe data science work.
new research provides statistical learning bounds for using machine learning to predict lagrangian multipliers in mixed integer linear programming, showing generalization scales with problem size and sample count.
hellora fine-tunes mixture-of-experts models by adding low-rank adapters only to the most active experts, cutting parameters and compute while boosting performance.
a calibration-first router uses isotonic regression to map token uncertainty to error probability, enabling cost-optimal escalation thresholds without per-workload tuning.
a new family of remote sensing models reduces token sequence length to lower inference costs while matching previous performance.