All authors
Claude Skills by hiyenwong
github.com/hiyenwong9,934 skills5 installs19,223 views
- Cbtr Causality Topological RankingCausality-Based Topological Ranking (CBTR)Votes: 0GitHub stars: 3
- Chart Visual ReasoningSkill for AI agent capabilitiesVotes: 0GitHub stars: 3
- Differentiable Clone Structured Causal GraphsSkill for implementing the differentiable Clone-Structured Causal Graph (gradCSCG) algorithm for end-to-end cognitive map learning from raw image sequences, as described in arXiv:2607.12382.Votes: 0GitHub stars: 3
- Differential Dynamic Causal NetsSkill for AI agent capabilitiesVotes: 0GitHub stars: 3
- Do Latent Channels Actually Communicate A Causal ADo Latent Channels Actually Communicate? A Causal Audit of Latent Multi-Agent LLMVotes: 0GitHub stars: 3
- Efficient Reasoning BcrReduce LLM reasoning token consumption using Batched Contextual Reinforcement (BCR). Use when optimizing inference costs for reasoning tasks, implementing efficient Chain-of-Thought, or discovering task-scaling laws. Based on arXiv:2604.02322 - A Task-Scaling Law for Efficient Reasoning.Votes: 0GitHub stars: 3
- Explainable Ai Xai SurveySkill for AI agent capabilitiesVotes: 0GitHub stars: 3
- Factorized Inference In Deep Markov Models For Incomplete Multimodal Time Series**arXiv ID:** 1905.13570 **Authors:** Tan Zhi-Xuan, Harold Soh, Desmond C. Ong **Published:** 2019-05-30T10:20:32Z **Abstract:** Integrating deep learning with latent state space models has the potential to yield temporal models that are powerful, yet tractable and interpretable. Unfortunately, current models are not designed to handle missing data or multiple data modalities, which are both prevalent in real-world data. In this work, we introduce a factorized inference method for Multimodal ...Votes: 0GitHub stars: 3
- Feature Leakage Identifiability Entropy ModelsIdentifiability framework for direct-dependency entropy models of neural activity - diagnosing feature leakage in MaxEnt models, separating prediction from mechanism identification, state reweighting diagnosticsVotes: 0GitHub stars: 3
- Func Lingam Causal DiscoverySkill for AI agent capabilitiesVotes: 0GitHub stars: 3
- Graph Laplacian J Divergence**来源论文：** arXiv:2012.11240 - Improving J-divergence of brain connectivity states by graph Laplacian denoisingVotes: 0GitHub stars: 3
- Hierarchical Bayesian Statistical Learning EegHierarchical Bayesian Statistical Learning (HBSL) model for individual statistical learning trajectories from EEG data. Models how individuals discover structure in sensory sequences, with applications to dyslexia research and cognitive development. Activation: hierarchical Bayesian, statistical learning, EEG, individual differences, dyslexia, sequence structure, tone sequences.Votes: 0GitHub stars: 3
- Improving Mathematical Reasoning With Process SupeSkill for AI agent capabilitiesVotes: 0GitHub stars: 3
- Indefinite Causal Order Real ComplexIndefinite causal order methodology demonstrating that real quantum theory with indefinite causal order can simulate complex quantum theory, reversing the conventional real-complex hierarchy.Votes: 0GitHub stars: 3
- Information Theoretic PirInformation-theoretic authenticated private information retrieval (aPIR) methodology. Use when: implementing privacy-preserving data retrieval, designing secure query protocols, building authenticated retrieval systems, or analyzing information-theoretic security guarantees. Covers unconditional security against malicious adversaries with information-theoretic privacy and authenticity guarantees. Keywords: private information retrieval, PIR, aPIR, information-theoretic security, privacy-prese...Votes: 0GitHub stars: 3
- Information Theoretic Portfolio SelectionPortfolio selection methodology using information projection and Renyi divergence decomposition under CRRA utility. Decomposes certainty-equivalent growth rate into portfolio-induced Renyi divergence, Renyi entropy of risk-tilted market law, and log-partition term. Use when designing portfolio selection strategies, applying information theory to finance, optimizing under risk aversion, or analyzing market payoff distributions through divergence measures.Votes: 0GitHub stars: 3
- Information Theory Neural Coding应用信息论框架分析神经编码和神经群体动力学。包含互信息、信息瓶颈、传递熵等方法在神经科学中的应用。Votes: 0GitHub stars: 3
- Knowledge Before Reasoning Ec Reason Bench A TrainKnowledge before Reasoning: EC-Reason-Bench, a Training-Free Diagnostic Benchmark for LLM Enzyme ClaVotes: 0GitHub stars: 3
- Majorana Fermion Topological GatesTopological quantum gate design using Majorana fermion motion methodology. Develops planar Pauli stabilizer codes and logical gate protocols via point-like Majorana fermions. Information stored in pairwise fermion parity, enabling fault-tolerant quantum computation through topological protection. Activation: Majorana fermion, topological quantum computing, Pauli stabilizer, logical gate design, quantum error correction, topological protectionVotes: 0GitHub stars: 3
- Majorization Entropy InequalitiesMajorization lattice framework for proving entropy inequalities in classical and quantum information theory. Covers supermodularity and subadditivity of all sum-concave entropies (Shannon, Rényi, Tsallis) via structural majorization relations. Use when analyzing entropy inequalities, information-theoretic bounds, quantum state entropy comparisons, or proving subadditivity/supermodularity results.Votes: 0GitHub stars: 3
- Maximum Entropy Neural ConnectivityMaximum entropy framework for deriving minimally-biased neural network connectivity that satisfies functional constraints for context-dependent computations. Reveals low-rank structures required for working memory, context integration, and task switching while keeping other connectivity aspects random.Votes: 0GitHub stars: 3
- Mcmc Hierarchical Dcm ClusteringMarkov chain Monte Carlo methods for hierarchical clustering of dynamic causal models, enabling subgroup detection in heterogeneous populations through effective brain connectivity analysis.Votes: 0GitHub stars: 3
- Modular State Space Model 260714078Model human perception, cognition, and decision dynamics as a modular perception-cognition-decision pipeline state-space model. Provides mathematical formulation, stability conditions, and application to rehabilitation control. Use when you need interpretable dynamical models linking neural mechanisms to behavior.Votes: 0GitHub stars: 3
- Multilingual ReasoningUnderstanding cross-language reasoning patterns in Large Reasoning Models. Use when designing multilingual reasoning systems, building non-English LLM applications, or addressing language-specific reasoning optimization. Triggers on "multilingual reasoning", "cross-language reasoning", "non-English reasoning models", or "language-specific reasoning patterns".Votes: 0GitHub stars: 3
- Native Explainability BcpnnFirst XAI taxonomy for BCPNN mapping architectural primitives to 16 explanation primitives (P1-P16) and 5 design-time Configuration-as-Explanation primitives. Inherently transparent brain-like neural network with EU AI Act compliance.Votes: 0GitHub stars: 3
- Nonlinear Cross Entropy BenchmarkingSample-efficient quantum advantage benchmarking using nonlinear cross-entropy and heavy output generation classifiers. Use when: (1) benchmarking NISQ quantum circuits, (2) distinguishing quantum computers from classical spoofers, (3) designing quantum advantage experiments, (4) analyzing random circuit sampling results, (5) evaluating shallow-depth quantum circuits.Votes: 0GitHub stars: 3
- Probabilistic Cbf SubgaussianProbabilistic Control Barrier Functions for safety-critical systems with state estimation uncertainty using sub-Gaussian concentration. Provides finite-sample safety certificates via particle-based CVaR estimation. Use for spacecraft proximity operations, safety-critical control under uncertainty, and formal safety guarantees with probabilistic constraints.Votes: 0GitHub stars: 3
- Probabilistic Compositional InferenceProbabilistic Compositional Inference methodology for coupled engineered systems - graph-based architecture for uncertainty-aware inverse inferenceVotes: 0GitHub stars: 3
- Probe Trajectory Reasoning MonitoringProbe trajectory methodology for monitoring Large Reasoning Model (LRM) internal dynamics. Tracks concept probability evolution across Chain of Thought tokens using signal-processing features (volatility, trend, steady-state) to predict future model behavior.Votes: 0GitHub stars: 3
- Reasoning As A Double Edged Sword Architecture AndDerived from arXiv:2607.17786 - Reasoning as a Double-Edged Sword: Architecture and Cross-Stage Robustness in Vision-Language-Action ModelsVotes: 0GitHub stars: 3
- Reasoning Decision TimingUnderstanding when LLM reasoning models make decisions - before or during chain-of-thought. Use when discussing reasoning model interpretability, AI safety, chain-of-thought reliability, or the philosophical implications of LLM decision-making processes. Triggers on questions about "reasoning models decide first", "chain-of-thought rationalization", "LLM interpretability", or "reasoning timing".Votes: 0GitHub stars: 3
- Reasoning Driven RetrievalRetrieval as iterative reasoning methodology. Treat retrieval as explicit hypothesis-driven search with evidence evaluation and self-improving refinement. Use when building RAG systems, information retrieval agents, search optimization, or any system that needs to go beyond black-box retrieval to find latent-pattern documents.Votes: 0GitHub stars: 3
- Stepwise Reasoning SubgraphStepwise reasoning framework that builds query-specific subgraphs from external knowledge bases to ground intermediate reasoning steps, improving LLM reasoning accuracy and factual reliability.Votes: 0GitHub stars: 3
- Training Continuous Chain Of Thought Models A TaleDerived from arXiv:2607.16972 - Training Continuous Chain of Thought Models: A Tale of Two RegimesVotes: 0GitHub stars: 3
- Vla Probabilistic Chunk MaskingDrop-in GRPO modification that allocates gradient computation to a small, probabilistically selected subset of trajectory chunks using success-failure action variance. Achieves 2.38x wall-clock speedup while matching final performance.Votes: 0GitHub stars: 3
- Whisperrec Latent Reasoning For Efficient FoundatiWhisperRec: Latent Reasoning for Efficient Foundation Recommendation ModelsVotes: 0GitHub stars: 3
- Advantage Collapse Grpo AvspoAdvantage Collapse in Group Relative Policy Optimization (GRPO): Diagnosis and Mitigation via Adaptive Virtual Sample Policy Optimization (AVSPO). Introduces the Advantage Collapse Rate (ACR) metric to diagnose training stagnation, and proposes AVSPO to inject virtual reward samples guided by real-time ACR monitoring. Use when: diagnosing GRPO training failures, improving LLM reasoning RL post-training, mitigating advantage collapse, ICML 2026 accepted. Activation: advantage collapse GRPO, AV...Votes: 0GitHub stars: 3
- Agpo Adaptive Group Policy OptimizationAGPO (Adaptive Group Policy Optimization) methodology — a critic-free refinement of GRPO that uses group-level statistics to adaptively control update magnitude and exploration. Uses a shared probe-derived statistical state to drive adaptive clipping (based on reward dispersion, skewness, probe entropy, policy entropy, KL drift) and bidirectional adaptive temperature sampling. Outperforms PPO/GRPO on 9 math/STEM benchmarks with Qwen2.5-14B. Use when: improving GRPO training stability, reducin...Votes: 0GitHub stars: 3
- Arms Automatic Reward Shaping MarlARMS (Automatic Reward-shaping in Multi-agent Systems) — self-supervised reward shaping for sparse-reward MARL with Nash equilibrium preservation guarantees.Votes: 0GitHub stars: 3
- Bi Nac Bilevel Rl Textual FeedbackBilevel Natural Language Actor-Critic (Bi-NAC) methodology — joint training of a critic to generate reward-improving textual feedback and an actor to exploit it, formulated as a Stackelberg bilevel program for RL with learnable textual feedback.Votes: 0GitHub stars: 3
- Certificate Guided Rl GeneralizationLogic-driven framework for evaluating reinforcement learning generalization using certificate-guided evaluationVotes: 0GitHub stars: 3
- Clipping Bottleneck NsrNear-boundary Stochastic Rescue (NSR) for stabilizing RLVR/GRPO training via stochastic recovery of clipped signalsVotes: 0GitHub stars: 3
- Conditional Equivalence Dpo RlhfProves DPO and RLHF are conditionally equivalent (not universally), identifies failure modes when the implicit assumption is violated, and proposes Constrained Preference Optimization (CPO) for provable alignment. 49-page theoretical work with geometric interpretation. Use when: analyzing DPO vs RLHF trade-offs, building preference optimization systems, theoretical analysis of alignment algorithms. Activation: DPO RLHF equivalence, conditional equivalence, CPO, preference optimization theory,...Votes: 0GitHub stars: 3
- Curverl Distribution Aware RlvrCurveRL methodology — principled distribution-aware context reweighting for RLVR, using quantile coordinate transform where prompt weights depend on pass-rate rank and density rather than absolute values.Votes: 0GitHub stars: 3
- D2evo Dual Difficulty Self EvolutionD²Evo methodology — Dual Difficulty-Aware Self-Evolution for data-efficient reinforcement learning in LLM reasoning. Addresses Effective Data Scarcity and Dynamic Difficulty Shifts by automatically selecting medium-difficulty samples via dual difficulty scoring (performance-based + entropy-based). Use when: data-efficient RL post-training for LLMs, curriculum-free self-evolution, difficulty-aware sample selection, GRPO data optimization, RL training data management. Activation: D2Evo, dual di...Votes: 0GitHub stars: 3
- Daca Grpo Denoising Credit AssignmentDenoising-Aware Credit Assignment for GRPO in Diffusion Language Models. Introduces Denoising Progress Scores and Stratified Masking Likelihood to improve GRPO-style training for diffusion LLMs, achieving gains up to 5.6pp on math reasoning, 7.4pp on code generation, and 36.3pp on constraint satisfaction.Votes: 0GitHub stars: 3
- Delta Discriminative Token Credit AssignmentDelTA (Discriminative Token Credit Assignment) methodology for Reinforcement Learning from Verifiable Rewards (RLVR). Introduces a discriminator view of RLVR updates showing policy-gradient implicitly acts as a linear discriminator over token-gradient vectors. Proposes token-level coefficient estimation to amplify discriminative directions and downweight shared patterns (e.g. formatting tokens). Outperforms baselines by 3.26 pts on Qwen3-8B and 2.62 pts on Qwen3-14B across math benchmarks. Us...Votes: 0GitHub stars: 3
- Diffusion Marl Motion PlanningGenerative multi-robot motion planning using diffusion modeling with Multi-Agent Reinforcement Learning guidanceVotes: 0GitHub stars: 3
- Drpo Drifting Preference OptimizationDrifting Preference Optimization (DrPO) - Online preference-finetuning for one-step generative models using non-parametric dipole preference fields and reference drift estimation.Votes: 0GitHub stars: 3
- Efficient TdmpcEfficientTDMPC improves model-based RL for continuous control with ensemble dynamics, uncertainty-penalized planning, and data freshness optimizations. Achieves SOTA sample efficiency on HumanoidBench-Hard and DMC hard, with benefits from higher update-to-data ratios.Votes: 0GitHub stars: 3