The Archive

MCGenie generates standardized model cards from repository metadata; MCTidy reorganizes existing cards for consistency.

StarHarness evolves agent harnesses (prompts, tools, MCP providers, subagent structure) without retraining model weights.

LION applies Clifford algebra to multimodal-attributed graph neural networks, improving representation and downstream task performance.

CAFE co-evolves search agent and feedback critic to improve intermediate trajectory corrections without waiting for terminal rewards.

Perturbation audit reveals medical chain-of-thought rationales often don't drive model outputs; proposes clinically-motivated failure mode classification.

Linear probes achieve robust, sample-efficient detection of machine-generated text with better OOD generalization than supervised baselines.

StepGuard audits agent tool invocations step-by-step; introduces StepGen to auto-generate safe/unsafe trajectories for guardrail training.

Ensemble CNN system predicts stroke risk from 11 features using seven supervised ML algorithms on medical imaging data.

Approximates Bayesian Ideal Observer performance bounds for detection tasks via denoising score matching, avoiding MCMC resampling per image.

AtlasNav identifies Evidence Blindness in LLM agents using direct corpus interaction; proposes navigation strategy to surface and expose critical documents.

MoTE uses task-level sparse mixture-of-experts for multi-task video understanding, improving over dense decoders for action recognition and forecasting.

Google Search tips for home decor shopping and DIY projects; consumer lifestyle content unrelated to frontier AI.

FactoMap models disentangled representations with factor-space structure (domains, identifications, anisotropy) beyond Euclidean product assumptions.

Framework for generating CAD programs from natural-language descriptions by inferring missing spatial constraints using design experience.

RACE: Statistical framework for evaluating functional consistency of Transformer neurons at scale via forward-pass perturbation analysis.

Weakly supervised semantic segmentation for seagrass habitat mapping in side-scan sonar imagery using image-level labels.

RAT: Bayesian framework for evaluating RAG systems by jointly modeling retrieval, abstention, and answer correctness across pipeline components.

Sociological study of AI sensemaking via text analysis of millions of articles and interviews with 57 AI professionals in 2021–2023.

SkillForge: Framework for continuous skill evolution in RL agents via verification and refinement of reusable knowledge across episodes.

DNN²: Doubly non-negative relaxation for ReLU neural network verification using Burer-Monteiro factorization to close safety guarantee gaps.

Hardware-realized Turing Machine with autonomous multi-step execution and reprogrammable optical input for extended computations.

Meta^n: Recursive self-improvement framework for LLM agents that increases meta-depth beyond two by applying fixed meta-operation to its own outputs.

Minimax-optimal alternating regret bounds for online linear and convex optimization with O(log d) constant regret independent of horizon T.

Parameter-efficient adaptation of EEG foundation models via self-supervised learning using only 9% of parameters for clinical transfer tasks.

Arc-length velocity polygonal synthesis extends audio generation to arbitrary shapes, morphing, and 3D polyhedra via unified DSP pipeline.

KENDO framework integrates ensemble Gaussian processes with disagreement-aware acquisition for Bayesian optimization and active learning without MCMC.

Deep learning super-resolution methods downscale SEVIRI cloud masks and enable cross-sensor cloud mask translation for satellite data.

Lifted inference under approximate commutativity identifies and handles deviation in learned parameters for scalable probabilistic inference.

Constrained hyperparameter optimization for streaming/online learning with bounded search spaces and self-adjusting parameter tuning.

Multilingual spoken hallucination benchmark with 12k samples across English, Russian, Kazakh in text and audio with controlled hallucination types.