Menu
Skills in this repository
SkillsMP has collected 4,114 skills from hiyenwong/ai_collection. Open a skill to review its source and details.
hiyenwong/ai_collectionShowing 40 of 4,114 collected skills.
EdgeSpike: SNN framework for low-power autonomous sensing on edge IoT. Covers hybrid surrogate-gradient training, hardware-aware NAS, event-driven runtime for Loihi 2/SpiNNaker 2/ARM Cortex-M, and local plasticity for on-device adaptation. Activation: edge…
Cross-population framework for evaluating robustness and generalizability of EEG biomarkers in multi-site clinical settings. Addresses cross-subject and cross-platform variation for reliable Parkinson's disease detection. Keywords: EEG biomarkers,…
EEG脑连接BCI分析方法论。通过功能连接分析理解脑网络在BCI中的机制，用于神经康复和外骨骼控制。适用于脑机接口、神经康复、步态训练。触发词：EEG、脑连接、BCI、脑机接口、功能网络、神经康复、brain-computer interface、neurorehabilitation。
Source text: Chinese
Systematic benchmark of channel adaptation methods for EEG foundation models. Compares Conv1d, SSI, source-space decomposition, and Riemannian re-centering across 5 FMs (5M-157M params), 5 tasks, revealing architecture-dependent optimal methods and probe-SFT…
Deep Sleep Classification via EEG Signal Criticality using Detrended Fluctuation Analysis (DFA) for passive Brain-Computer Interface (pBCI) neurofeedback applications. Probabilistic decoding of EEG criticality features for state-dependent sleep improvement…
Structure-Guided Diffusion Model (SGDM) for EEG-based visual cognition reconstruction. Combines structurally supervised VAE, spatiotemporal EEG encoder with contrastive learning, and ControlNet-guided diffusion for high-fidelity visual reconstruction from…
EEG-fused digital twin brain framework for autonomous driving in virtual scenarios. Combines biophysical brain models with EEG data and digital twin technology for driver state monitoring and vehicle control. Applies to: brain-computer interfaces, autonomous…
EEG-based emergency braking intensity prediction using blind source separation. Artifact removal for reliable EEG-based driver assistance systems. Activation: eeg braking, blind source separation, driver assistance, artifact removal.
Fast Automatic Artifact Rejection (FAAR) methodology for EEG motor imagery BCIs. Lightweight automated artifact rejection that computes artifact-sensitive features, derives epoch-level Signal Quality Index, adaptively selects rejection thresholds, and…
EEG基础模型系统评估和分析管道。提出ASHA基准测试、范式级消融研究、神经生理学探测(NPP)框架，确保EEG基础模型的公平评估和可解释性。
Source text: Chinese
EEG-conditioned framework for reconstructing dynamic fMRI as continuous neural sequences with high spatial fidelity and temporal coherence at cortical-vertex level. Incorporates null-space intermediate-frame reconstruction for handling sampling irregularities.
Layer-wise Relevance Propagation (LRP) methodology for interpreting EEG foundation models. Extends LRP from CNN-based to Transformer-based EEG models, enabling verification and hypothesis discovery. Activation: EEG interpretability, LRP, EEG foundation model,…
EEG foundation models with domain adaptation using lightweight adapters. Covers pre-trained EEG encoders, task-specific fine-tuning with adapters, cross-dataset generalization, and efficient deployment. Use when working with EEG foundation models, neural…
Mechanistic interpretability of EEG foundation models using Sparse Autoencoders (SAEs). Extracts interpretable feature dictionaries from EEG transformer embeddings via TopK SAEs, benchmarks monosemanticity across architectures (SleepFM, REVE, LaBraM), and…
EEG-based Hopfield energy landscape analysis for quantifying brain network stability during emotional processing (happy/sad face tasks). Activation: emotion energy landscape, brain stability, happy sad face EEG, Hopfield emotion.
Energy landscapes for quantifying brain network stability during emotional processing. Uses Hopfield network energy framework to analyze EEG dynamics, mapping emotional states to attractor basins in brain network energy landscapes. Provides a physics-based…
Bridging scalp EEG and intracranial EEG (iEEG) in BCI via pretrained neural models. Maps non-invasive scalp EEG to iEEG-quality representations, enabling high-fidelity BCI without invasive implants. Uses pretrained models to learn the scalp-to-cortical…
EEG-MFTNet: Enhanced EEGNet with multi-scale temporal convolution and fusion transformer for cross-session motor imagery decoding. Addresses session variability in BCI through dual-branch architecture combining frequency-specific temporal features with global…
Universal EEG microstate tokenizer for representation learning across downstream tasks. Clusters continuous EEG into discrete microstate tokens that serve as universal building blocks for sleep staging, emotion recognition, seizure detection, and motor…
Interpretable EEG microstate discovery via variational deep embedding with systematic architecture search and multi-quadrant evaluation. Uses deep variational methods for data-driven microstate identification instead of traditional k-means clustering on GFP…
EEG preprocessing reliability methodology for quantifying and mitigating preprocessing-induced prediction instability in EEG deep learning. Based on arXiv:2605.07212 (Hou et al., 2026). Use when: (1) evaluating EEG model robustness to preprocessing pipeline…
Mechanistic interpretability of EEG foundation models via sparse autoencoders. Extracting interpretable features from EEG foundation model internal representations using sparse autoencoder decomposition. Use when: interpreting EEG foundation models,…
Research methodology and findings on using EEG cortical tracking strength (CTS) as a neural marker for early-stage cognitive decline. Combines speech encoding models with linguistic feature analysis to detect subjective cognitive decline (SCD). Use when:…
Subject-specific analysis of self-initiated attention shifts from EEG using interpretable machine learning. Demonstrates reliable within-subject classification of preparatory EEG activity distinguishing self-initiated vs externally instructed attention…
Neuroscience-inspired staged representation learning framework for EEG visual decoding. Organizes EEG representation learning into three complementary phases: low-level visual, high-level semantic, and integrative fusion, with disentangled coarse/fine-grained…
Structure-Guided Diffusion Model (SGDM) for EEG-based visual reconstruction. 通过结构引导的扩散模型实现从脑电信号到视觉图像的重建。
Source text: Chinese
Structure-Guided Diffusion Model (SGDM v2) for EEG-based visual cognition reconstruction with enhanced cross-subject generalization. Activation: EEG diffusion reconstruction, visual cognition decoding, SGDM, brain-to-image, neural decoding.
Structure-Guided Diffusion Model (SGDM v3) for EEG-based visual cognition reconstruction with enhanced cross-subject generalization. Combines structurally supervised VAE, spatiotemporal EEG encoder with contrastive learning, and ControlNet-guided diffusion…
Structure-Guided Diffusion Model (SGDM v4) for EEG-Based Visual Cognition Reconstruction. Diffusion-based framework for reconstructing visual stimuli from EEG with structural guidance for improved accuracy. Activation: SGDM, EEG reconstruction, visual…
Deep learning framework for objective consciousness level measurement using multi-dimensional transcranial electrical stimulation (TES) with EEG. Combines TES-evoked brain responses with CNN classification for bedside-awareness assessment. Activation…
NeuroAdapt-Bench: Systematic benchmark for test-time adaptation (TTA) on EEG foundation models under real-world distribution shifts. Evaluates TTA methods across multiple FMs, tasks, and datasets including extreme modality shifts (Ear-EEG). Finds…
EEG-based tinnitus biomarker identification methodology with cross-dataset generalization. Uses microstate analysis and Koopman operator analysis via DMD to extract robust neural signatures. Focuses on Koopman eigenvalue magnitude for oscillation stability.…
Benchmarking positional encoding strategies for transformer-based EEG foundation models. Systematic evaluation of five positional encoding strategies within CBraMod backbone for motor imagery classification and emotion recognition. Key findings: SPE excels at…
EEG-based visual attention decoding from gaze-fixated neural tracking of motion in natural videos. Addresses eccentricity confounds and eye movement artifacts for brain-computer interface research. Activation: EEG attention decoding, visual attention BCI,…
EEG2Vision — Modular end-to-end EEG-to-image reconstruction framework using diffusion models with MLLM-guided boosting. Evaluates performance across EEG resolutions (128/64/32/24 channels). Enables real-time brain-to-image applications with low-density EEG.…
Network-based framework for quantifying plasticity as system_size/connectivity_ratio. Defines effective plasticity as normalized measure linked to critical regime. Plasticity drives criticality causally. Activation: effective plasticity, plasticity…
Effective rank methodology for predicting quantum data encoding performance. Uses feature map effective rank as a threshold criterion to accelerate the search for high-performing QML encodings. Activation: effective rank encoding, feature map rank QML,…
Efficient Clifford+T synthesis methodology for small-angle rotations with application to Trotterization - reducing T gate cost from O(log 1/δ) to Õ(θ²/δ) for small angles in fault-tolerant quantum compilation.
Theoretical framework linking efficient coding to criticality in neural populations. Shows that maximizing Fisher information under resource constraints naturally leads to soft modes, diverging correlation lengths, and power-law neural avalanches, unifying…
Efficient coding under resource constraints drives neural systems towards criticality and sloppiness. Links Fisher information maximization to power-law distributions and critical brain hypothesis.