菜单
这个仓库中的 skills
SkillsMP 已收集 hiyenwong/ai_collection 中的 4,114 个 Skill。打开任一 Skill 可查看来源和详情。
hiyenwong/ai_collection已展示 40 / 4,114 个已收集 Skill。
Decentralized Momentum Tracking with Biased Gradients (Biased-DMT) for large-scale distributed optimization. Handles communication compression and data heterogeneity in decentralized learning. Use for: decentralized optimization, federated learning,…
原文语言：英语
声明式自我改进技能，通过目标驱动的方式引导 Agent 持续进化。触发词：自我改进、self-improvement、目标驱动、goal-driven improvement、声明式进化、declarative evolution、改进循环、improvement loop。
Complexity-theoretic benchmarking methodology for decoded quantum interferometry (DQI) and bounded-degree constraint satisfaction problems. Analyzes quantum advantage limits, classical vs quantum decoder performance, and approximation hardness over finite…
原文语言：英语
Critical analysis framework for brain-model alignment methodology. Demonstrates that representational similarity analysis (RSA) and decoding-based alignment metrics are fundamentally insensitive to encoding manifold topology. Similar decoding behavior and…
原文语言：英语
深度连续局部学习方法论（DECOLLE）。在脉冲神经网络中实现局部突触可塑性规则，通过合成梯度实现端到端训练。适用于事件驱动视觉、神经形态计算、在线学习、脉冲神经网络研究。触发词：DECOLLE、脉冲神经网络、突触可塑性、局部学习、神经形态计算、在线学习、spiking neural network、synaptic plasticity、neuromorphic computing。
Distance coding via de-correlation of heterogeneous grid cell populations. Mathematical theory showing how small variability in grid properties enables distance encoding through population activity de-correlation, with non-intuitive 'sweet spot' predictions…
原文语言：英语
Deep binarized photonic reservoir computing architecture achieving Gb/s multimedia signal processing via digital micro-mirror device (DMD), optical scattering, and CMOS photodetection. Use for ultra-fast video/image/speech recognition, neuromorphic photonic…
原文语言：英语
Network defensibility analysis using shield synthesis and adversarial game theory. Reinterprets shielded RL from runtime enforcement to design-time structural analysis. Use when: analyzing network topology security, synthesizing safety shields for RL agents,…
原文语言：英语
延迟自适应脉冲神经网络分类器。基于共形预测(CP)提供可靠性保证的早停机制，让SNN在足够自信时提前决策，降低延迟和能耗。适用于神经形态计算、边缘AI、实时推理。触发词：SNN早停、延迟自适应、共形预测、脉冲神经网络、可靠性保证、delay-adaptive、early stopping、conformal prediction、spiking neural network。
Differential structural connectivity analysis in dementia and aging using the OASIS-3 dataset. Reveals both decreased connectivity (hippocampus, temporal lobe) and surprisingly increased connectivity (precuneus, cuneus, insula) in demented brains. Activation:…
原文语言：英语
DendroNN dendrocentric neural network methodology for energy-efficient classification of event-based data. Incorporates dendritic computation principles into SNNs for improved spatiotemporal processing. Applies to: event-based vision, neuromorphic computing,…
原文语言：英语
Optimal decoding methodology for quantum error correction using density matrix propagation through circuit-level noise. Enables ML-decoding benchmarking with pruning techniques and rigorous bounds for repetition codes and cellular automaton codes. Activation:…
原文语言：英语
Dependability engineering for hybrid quantum-classical computing systems. Covers reliability, resiliency, security, and reproducibility patterns for quantum hardware integration with classical HPC infrastructure. Includes fault-tolerance verification, error…
原文语言：英语
Description Length (DL) and Fractional Bayes Factor (FBF) model selection methodology for genetic programming and symbolic regression. Evaluates DL/FBF as principled alternatives to AIC/BIC for selecting compact, generalizable expressions. Use when working…
原文语言：英语
Methodology for defending AI agents against prompt injection using constraint-based response limiting and defense in depth. Based on OpenAI's agent security approach achieving 42% reduction in injection attempts. Use when building secure agent systems,…
原文语言：英语
Evaluation methodology for detecting hidden misalignment ("scheming") in AI models and concrete methods for reducing deceptive behavior.
原文语言：英语
Deterministic generation of large cat states (100+ photons) using dynamical invariants of hybrid qubit-bosonic systems under time-dependent Hamiltonians. Universal quantum control theory for quantum metrology and fault-tolerant computation. arXiv:2606.03293.
原文语言：英语
Developmental neural circuit generation methodology from gene regulatory rules. Simulates cortical neurogenesis from single stem cell to generate domain-general topological substrates amenable to rapid learning. Use when studying developmental priors for…
原文语言：英语
Quantum reservoir computing using decoherence-free subspaces (DFS) for room-temperature quantum AI. Classifies entangled vs product states without cooling. Activation: quantum reservoir, decoherence-free subspace, DFS, room temperature quantum, quantum…
原文语言：英语
DGCL Brain Network Construction
原文语言：英语
Dynamic Gated Neuron (DGN) — brain-inspired gating mechanism for robust spiking neural networks. Dynamic conductance acts as biologically plausible gating for selective input filtering and adaptive noise suppression. Use when designing robust SNNs,…
Diagonal Adaptive Non-local Observables (ANO) methodology for quantum neural networks — reduces observable parameter complexity from O(4^k) to O(2^k) while retaining full ANO expressivity via diagonal canonical representatives. For efficient VQA measurement…
原文语言：英语
Differentiable biophysical simulation framework for inferring Hodgkin-Huxley parameters from extracellular MEA data. Enables rapid biophysical inference and precise neurostimulation prediction without invasive intracellular recordings.
原文语言：英语
Differentiable physical framework for goal-driven spin-state engineering in Magnetic Resonance Spectroscopy (MRS). Use when: designing MRS pulse sequences via automatic differentiation, navigating high-dimensional spin dynamics, engineering complex quantum…
原文语言：英语
Error Diffusion (ED) methodology for biologically plausible credit assignment under Dale's principle. Dual-stream excitatory/inhibitory architecture with modulo error routing. Achieves 96.7% MNIST and 61.7% CIFAR-10 under strict Dale's constraint. Integrates…
原文语言：英语
Inferring active neural circuits from neural activity data using diffusion scores methodology. Identifies functionally connected neural populations through diffusive signal propagation patterns. Activation: diffusion scores, neural circuit inference, active…
原文语言：英语
Optimal sequential decision-making for error propagation mitigation in modular digital twins using MDP/POMDP framework. Combines HMM-based latent regime inference with MDP/POMDP intervention selection. Features: data-driven transition model, Point-Based Value…
原文语言：英语
DINA (Dual-Tower Image-Neural Alignment) framework for interpretable contrastive analysis of V1 population activity. Aligns visual stimuli and V1 responses in shared latent space at intermediate feature map level. Activation: DINA, V1 population activity,…
原文语言：英语
Reversible quantum oracle construction for solving bounded Diophantine systems via amplitude amplification. Use when designing quantum algorithms for integer optimization, constraint satisfaction over bounded domains, or synthesizing arithmetic circuits for…
原文语言：英语
DIRECT (DIRectional Edge Coupling/Training) methodology for causal learning with neural assemblies using local plasticity. Enables neural assemblies to internalize causal directionality without backpropagation. Activation triggers: neural assemblies, causal…
原文语言：英语
Direct-to-Event (D2E) Spiking Neural Network Transfer methodology. Converts direct-coded SNNs trained with floating-point inputs into energy-efficient event-based representations using Self-Knowledge Distillation (SKD). Activation: SNN transfer,…
原文语言：英语
Directional coordination hierarchy in large-scale brain dynamics — identifies three recurrent resting-state coordination regimes (feedback-dominated, feedforward-dominated, integrative), shows this framework is disrupted in schizophrenia, and links…
原文语言：英语
DisAgg protocol for efficient secure aggregation in federated learning using distributed aggregator committees, eliminating homomorphic encryption overhead while preserving privacy.
原文语言：英语
Model Predictive Control (MPC) stability and suboptimality analysis under plant-model mismatch with discounting. Provides theoretical guarantees for infinite-horizon optimal control when using surrogate models. Use when designing robust control systems,…
原文语言：英语
Discounted Model Predictive Control (MPC) and infinite-horizon optimal control under plant-model mismatch. Unified framework for stability and suboptimality analysis with robustness guarantees. Use for: robust MPC, plant-model mismatch handling, discounted…
Discrete Heat Kernels on Simplicial Complexes and Its Application to Functional Brain Networks. Unified framework for heat kernel smoothing on simplicial complexes extending classical signal processing methods to higher-order network structures.
原文语言：英语
Discrete signaling mediates chaotic regularization in recurrent neural networks - theoretical framework linking microscopic chaos to macroscopic geometry of neural representations. Activation: chaotic regularization, discrete signaling, RNN chaos, neural…
原文语言：英语
Dissipative quantum chaos methodology — extending Hamiltonian quantum chaos to open quantum dynamics via Lindbladian spectral analysis. Use when analyzing chaoticity of open quantum systems, distinguishing integrable from chaotic dynamics, or studying…
原文语言：英语
Distributed Hierarchical Temporal Memory (D-HTM) neuromorphic framework enabling cross-entity preemptive warning via Shared Associative Memory (SAM) — extends HTM beyond reactive detection to distributed predictive reasoning (arXiv: 2606.31789)
原文语言：英语
Communication-efficient distributed Inverse Quantum Fourier Transform (iQFT) using communication horizon pruning to reduce inter-node quantum communication from O(P^2) to O(P). Use when: distributed quantum computing, iQFT across quantum networks, quantum…
原文语言：英语