The Archive

Earphone-based passive biometric authentication system using in-ear accelerometers for heartbeat identification, unrelated to AI frontier topics.

Distance-metric-based instance methods detect conditional anomalies in patient management alerts.

Joint spectral radius analysis provides convergence guarantees for deflated Q-value iteration in discounted Markov decision processes.

An interactive visualisation of Jensen–Shannon divergence - the symmetric, always-finite cousin of KL. Shape two distributions and watch JSD, its ceiling of one bit, and the per-point contribution respond in real time…

Hermes Agent surpassed Claude Code and OpenClaw as top token consumer on OpenRouter in past 24 hours.

Reddit post claims Claude Mythos exceeded METR's time-horizon benchmark; lacks verification and uses hyperbolic framing.

Reddit post claims early Claude Mythos preview with 17-hour median time-to-completion metric; lacks detail, verification, or official source.

Temporal knowledge drift in LLMs is orthogonal to correctness in representation space; linear probes achieve 0.83–0.95 AUROC.

Dynamic Meta-Metrics proposes source-sentence conditioned weighting for machine translation evaluation across language pairs.

METR evaluated Claude Mythos Preview on autonomous task completion, estimating 16hr median time-horizon with methodological limitations at the measurement ceiling.

I built a small interactive explorer for building intuition about KL divergence: https://robotchinwag.com/posts/kl-divergence-visualisation/ You control two skew-normal distributions and can see the KL integrand and the KL metric. It’s good for exploring how it changes with a…

Proposes DR-ME, a semiparametrically efficient test for detecting distributional treatment effects beyond means in causal inference.

FLAM framework for computing aggregatable evaluation metrics in federated learning to reconcile local and global performance assessment.

Community critique: LLM benchmarks should include realistic context sizes, multimodal feature usage, and agentic/RAG workloads rather than speed-only metrics.

Improves geometric representations for molecule generative models by decoupling representation learning from 3D structure generation.

FactoryBench evaluates LLMs and time-series models on industrial robotic telemetry using Pearl's causal ladder and LLM-as-judge scoring.

Geometric framework analyzes directional accuracy of low-precision number representations in ML vector operations.

OpenAI documents sandboxing, approvals, network policies, and telemetry for safe Codex deployment in agent workflows.

Bharat ABIS describes billion-scale multimodal biometric search system for national identity using fingerprint, face, iris matching.

Benchmark evaluates latest LLMs on grammatical error correction across edit precision, fluency, and meaning retention with reference-free metrics.

SAM 3D Animal: multi-animal 3D reconstruction from single images using SMAL+ parametric model and prompt-based disambiguation.

Lightweight asymmetric neural codec for compressing multi-modal sensor data on bandwidth-constrained edge devices.

Formalization of Algospeak coevolution dynamics between LLM evasion and detection; introduces Majority Understandable Modulation metric.

Theoretical analysis connecting GNN encoder geometry to Graph Edit Distance estimation quality in neural graph matching.

STAT: coordination-aware evaluation framework for multi-agent RL, augmenting aggregate metrics with process-level diagnostic insights.

Reddit discussion on OpenAI's aggregate energy consumption metrics; data-focused, no ethical framing.

Reddit post urging opposition to GUARD Act, which would mandate ID/biometric verification for all AI chatbot access in the US.

Grok AI model discovered five new mathematical inequalities and bounds in convex geometry and combinatorics, verified by human authors.

Geometry-Aware State Space Model applies hyperbolic geometry to whole-slide histopathology image analysis via Multiple Instance Learning, improving patch aggregation for gigapixel resolution.

Manifold steering interventions causally link neural activation geometry to model behavior via structured representation space.