Menü
Skills in diesem Repository
SkillsMP hat 4.114 Skills aus hiyenwong/ai_collection gesammelt. Öffne einen Skill, um Quelle und Details zu prüfen.
hiyenwong/ai_collectionEs werden 40 von 4.114 gesammelten Skills angezeigt.
Investigates whether Large Reasoning Model (LRM) chain-of-thought reasoning effort (inference-time compute budget) aligns with human cognitive costs. Finds that cognitive cost alignment between LRMs and humans is a training-time achievement, robust to…
Quellsprache: Englisch
Matryoshka-style elastic Spiking Transformer with runtime-adaptive width and attention head slicing for deployment across hardware budgets without retraining. Reduces spike firing rates proportionally to parameter footprint. Use when deploying SNNs on…
Quellsprache: Englisch
Matryoshka-style elasticity for Spiking Transformers - runtime-adaptive architecture enabling dynamic width and attention head slicing at inference without retraining. Applies to SNN deployment on neuromorphic hardware, edge AI, and adaptive computation.…
Quellsprache: Englisch
Expert guidance for TypeScript, Electron, and Desktop App Development. Use when building desktop applications with Electron, working with IPC, or developing cross-platform desktop apps. Triggers on: electron, desktop app, ipc, main process, renderer process,…
Quellsprache: Englisch
ELSA — ELastic SNN Inference Architecture for efficient neuromorphic computing, featuring near-SRAM spine/token-wise dataflow pipeline, bundled AER protocol for NoC, and mini-batch spiking Gustavson-product for exploiting SNN sparsity. ISCA 2026. 3.4× speedup…
Quellsprache: Englisch
Feasibility analysis and hybrid architecture design for embedding quantum machine learning workloads in resource-constrained embedded systems. Explores the intersection of quantum computing and edge/embedded deployment. Use for: embedded quantum ML…
Quellsprache: Englisch
EMBER: Autonomous cognitive behaviour from learned spiking neural network dynamics. Self-organizing SNN agents with intrinsic motivation, curiosity, and goal-directed behavior emerging from plastic recurrent connectivity without external reward shaping.…
Quellsprache: Englisch
EMBER (Experience-Modulated Biologically-inspired Emergent Reasoning) - Hybrid cognitive architecture combining Spiking Neural Networks (SNN) with Large Language Models (LLM). SNN serves as persistent associative substrate, LLM as replaceable reasoning…
Quellsprache: Englisch
EMBER hybrid cognitive architecture combining LLM reasoning with persistent biologically-grounded SNN memory substrate. Autonomous cognitive behavior via STDP-based lateral propagation.
Quellsprache: Englisch
Hybrid LLM-SNN cognitive architecture that reorganizes the LLM-memory relationship by placing the LLM as a replaceable reasoning engine within a persistent, biologically-grounded SNN associative substrate. Features experience-modulated dynamics and emergent…
Quellsprache: Englisch
Embodied Neurocomputation framework for interfacing biological neural cultures with scaled task-driven validation. Systems-level approach to multi-variable optimization of encoding/decoding between silicon computing and living biology. Demonstrates that…
Quellsprache: Englisch
Embodied Neurocomputation framework for interfacing biological neural cultures (BNNs) with task-driven validation. Addresses the encoding/decoding optimization problem between silicon computing and living biological neural networks. Demonstrates first…
Quellsprache: Englisch
Embodied Virtual Reality feedback reshapes neural representations to support continuous 3D motor imagery decoding in brain-computer interfaces. First systematic investigation of embodied VR feedback during real-time 3D virtual limb control. Use when: (1)…
Quellsprache: Chinesisch
Embodied Virtual Reality feedback methodology for continuous 3D motor imagery BCI decoding. First systematic investigation showing VR feedback elicits more decodable and generalizable neural representations than screen feedback. CNN-LSTM decoder achieves…
Quellsprache: Englisch
涌现语言方法论 — 使用多智能体强化学习中的涌现语言作为研究意识的生成工具。从最小先验出发，通过任务压力发展通信，确保涌现结构的因果可归因性。
Quellsprache: Chinesisch
Emergent Systems Design - 自动化工程设计具有涌现属性的复杂系统。核心技术：描述性统计转损失函数、梯度下降优化涌现特征、Kuramoto耦合振荡器测试床。激活词：emergent design, emergence engineering, 涌现设计, 复杂系统工程.
Quellsprache: Chinesisch
End-to-end pipeline for decoding affective captions directly from fMRI signals using continuous emotion vectors and classifier-free guidance rewriting. Use when: brain-to-text with emotion/affect, fMRI affective decoding, continuous emotion representation…
Quellsprache: Englisch
EMRFormer: End-to-End Radar and Communication Modulation Recognition with Neuromorphic Computing using spike-driven transformers on KA200 neuromorphic chip. Achieves SOTA accuracy with 90%+ energy reduction. (arXiv: 2606.24075)
Quellsprache: Englisch
Energy-based Autoregressive Generation (EAG) framework for neural population dynamics using energy-based transformer in latent space with strictly proper scoring rules. Activation triggers: energy-based model, neural population dynamics, autoregressive…
Quellsprache: Englisch
Energy-based dynamical systems framework for neurocomputation, learning, and optimization. Unifies Hopfield networks, Boltzmann machines, modern EBMs, and equilibrium propagation under a single energy landscape formulation. Covers gradient flow dynamics,…
Quellsprache: Englisch
Energy-based regularization for learning residual dynamics in Neural MPC for omnidirectional aerial robots. Use when: (1) Designing neural network dynamics models for physical systems, (2) Implementing Model Predictive Control with learned dynamics, (3)…
Quellsprache: Englisch
Ensemble engineering methodology to overcome destructive cancellation in quantum measurements on NISQ devices. Addresses near-uniform ensemble sampling issues that render physically relevant expectation values unobservable. Activation: ensemble engineering,…
Quellsprache: Englisch
Unified physical theory linking entropy production, subjective time perception, and deformed neural dynamics. Derives conformable time operators from thermodynamic principles and predicts psychedelic time dilation and cognitive aging effects.
Quellsprache: Englisch
使用熵测度识别脑连接路径的方法论。通过信息论工具（熵密度、有效测度复杂度、Lempel-Ziv距离）检测线性与非线性动态，无需预设参数或模型假设。适用于任务态fMRI分析、脑区连接发现、探索性研究。触发词：脑连接、熵测度、信息流、fMRI分析、非线性动态、brain connectivity、entropy、information flow、Lempel-Ziv。
Quellsprache: Chinesisch
Internally triggered retrospective learning paradigm for neural networks. Instead of continuous externally-driven weight updates, parameter modifications are governed by internally generated events from the network's own representational dynamics. Uses latent…
Quellsprache: Englisch
Entropy Pacing Policy Optimization (EPPO) methodology for multi-task agentic RL. Coordinates entropy across tasks using dynamic clipping to prevent exploration-exploitation pace mismatch between tasks.
Quellsprache: Englisch
Equation-free digital twin framework using Koopman operator theory and Hankel-matrix embeddings for real-time structural state reconstruction without physical models. Use when: (1) building digital twins for complex engineering structures, (2) virtual sensing…
Quellsprache: Englisch
Equilibrium dynamics framework for microsecond-precision sound localization without explicit delay lines
Quellsprache: Englisch
Equilibrium Propagation (EP) with Predictive Learning in Leaky Integrate-and-Fire Spiking Neural Networks. Biologically plausible alternative to backpropagation for training SNNs using energy-based two-phase learning. Use when training SNNs without…
Quellsprache: Englisch
Equivariant Neural Belief Propagation (ENBP) for SE(3)-symmetric probabilistic inference with Gaussian mixture messages
Quellsprache: Englisch
Equivariant QAOA methodology incorporating symmetry constraints into quantum approximate optimization. Uses group-theoretic structure to reduce parameter space and improve optimization efficiency for combinatorial problems with inherent symmetries. Use when…
Quellsprache: Englisch
Equivariant reinforcement learning for Clifford quantum circuit synthesis. Use when synthesizing Clifford quantum circuits with RL, designing equivariant neural networks for quantum tasks, building size-agnostic policies across qubit counts, or optimizing…
Quellsprache: Englisch
Equivariant reinforcement learning for Clifford quantum circuit synthesis. Use when designing RL-based quantum circuit synthesis, leveraging group symmetries in quantum operations, or building equivariant architectures for quantum computing tasks.
Quellsprache: Englisch
ERA (Entropy-guided Visual Token Pruning with Rectified Attention) for efficient Multimodal Large Language Models
Quellsprache: Englisch
E-ReCON energy- and resource-efficient precision-configurable sparse nvCIM macro for conventional and spiking neural edge inference. Activation: nvCIM, ReRAM CIM, SNN hardware accelerator, edge-AI hardware, compute-in-memory SNN, neuromorphic hardware macro
Quellsprache: Englisch
Event-driven eligibility propagation (e-prop) extension for large sparse recurrent spiking networks. Biologically plausible learning rule with continuous dynamics, strict locality, and sparse connectivity. Scales to millions of neurons without compromising…
Quellsprache: Englisch
Event-driven asynchronous retrieval in high-capacity kernel Hopfield networks. KLR Hopfield networks achieve P/N ≈ 30 storage capacity with asynchronous updates, enabling energy-efficient neuromorphic deployment. Event count matches initial Hamming distance —…
Quellsprache: Englisch
Event-driven impulse radio transceiver system for reliable wireless neuromorphic inference. Ultra-low power event-based communication optimized for spike-based neural network data transmission. Triggers: event-driven radio, neuromorphic transceiver, impulse…
Quellsprache: Englisch
Event2Vec: Processing neuromorphic events directly via vector representations for efficient event camera data processing compatible with Transformer architectures. Activation triggers: event camera, neuromorphic vision, event2vec, DVS, asynchronous events,…
Quellsprache: Englisch
Evolutionary feature selection for spiking neural network pattern classifiers using the biologically realistic JASTAP model. Combines evolutionary algorithms with SNN training for simultaneous architecture and feature optimization.
Quellsprache: Englisch