Menu
Skills dans ce dépôt
SkillsMP a collecté 4 114 skills depuis hiyenwong/ai_collection. Ouvrez un skill pour examiner sa source et ses détails.
hiyenwong/ai_collectionAffichage de 40 skills collectés sur 4 114.
Anthropic research (Jun 8, 2026) — Measuring how LLMs dramatically accelerate N-day exploit development; Claude Mythos Preview built 8 working Firefox exploits autonomously and 8 Windows kernel privilege escalation chains, collapsing the historically slow…
Langue du texte source : anglais
Patterns for building AI-based medical diagnosis systems with clinical explainability. Covers foundation models for medical imaging, clinical reasoning trace generation, multi-modal patient data integration, and explainable AI for healthcare. Use when…
Langue du texte source : anglais
MemoryVLA++ - Temporal modeling framework for VLA models enabling persistent memory for long-horizon robotic manipulation tasks
Langue du texte source : anglais
MemoryWAM introduces persistent memory mechanisms for efficient world-action modeling with world model integration and hippocampal-inspired memory consolidation.
Langue du texte source : anglais
Self-organized learning in oscillatory neural networks (ONNs) using memristive signed couplings. Inhibitory (negative) weights enable anti-phase attractors, expanding accessible attractor structures beyond purely synchronous couplings for autonomous…
Langue du texte source : chinois
Reservoir computing with memristor dynamics for image classification. Studies how memristor intrinsic dynamics reduce network size and parameter overhead in reservoir computing for time-series prediction and image recognition. Trigger words: memristor…
Langue du texte source : anglais
Memristor-based spiking neural network accelerator for bio-inspired interception tasks - achieving 12.7x energy reduction vs digital SNN (arXiv:2605.31299v1, May 2026).
Langue du texte source : anglais
Mental fatigue-induced balance disturbance analysis using clustering-based heterogeneity classification. Investigating individual differences in balance control response to cognitive fatigue through AX-CPT and PVT performance metrics. Activation: mental…
Langue du texte source : anglais
梅森数与倍角映射的动力学联系研究。通过角度倍角映射动力学框架，无需显式计算M(n)即可求梅森数的因子。提供替代Lucas-Lehmer检验的动力学方法证明大梅森数为合数。适用于大数素性检验、动力系统数论应用。
Langue du texte source : chinois
脑组织中介尺度结构识别方法论。通过多分辨率技术提取脑连接网络的内聚结构。适用于脑网络模块化、社区检测。触发词：中介尺度、脑组织、模块化、mesoscale、community detection。
Langue du texte source : chinois
Meta-cognitive reflection framework for self-improvement based on reasoning review, error analysis, and learning strategy adjustment.
Langue du texte source : Plusieurs langues
Meta-cognitive framework for optimizing tool use in agentic multimodal models - deliberate tool invocation vs internal reasoning arbitration. Use when designing agents that need to decide between using external tools or internal knowledge. Activation:…
Langue du texte source : chinois
Meta-Representational Predictive Coding (MPC) — encoder-only neuroscience-informed self-supervised learning within the free energy principle, using cross-stream latent prediction and active inference saccade planning instead of backpropagation (arXiv:…
Langue du texte source : anglais
Metabolic quantum limit to the information capacity of magnetoencephalography - 代谢量子极限作为MEG信息容量的基本约束
Langue du texte source : chinois
Metabolic quantum limit methodology for magnetoencephalography (MEG) — derives technology-independent bounds on brain imaging information capacity using quantum sensing limits and neural metabolism.
Langue du texte source : anglais
"Metacognition-as-Reward (MaR) — metacognition-inspired RL framework for LLM reasoning. Use when training LLMs to reason better through RL: (1) improving reasoning quality beyond final-answer correctness, (2) providing reward signals for intermediate…
Langue du texte source : anglais
Physics-based metamorphic testing framework for Variational Quantum Circuits (VQCs). Addresses the oracle problem in quantum testing by deriving test oracles from quantum mechanical properties. Use when: testing VQEs/QAOA circuits, verifying quantum circuit…
Langue du texte source : anglais
Metastable Mind framework synthesizing Event Segmentation (ES) and Metastable Neural Activity (MNA) theories. Neural states as fundamental computational units with spatio-temporally nested hierarchy, predictive models, and modular processing boundaries.…
Langue du texte source : anglais
Metastable neural states as fundamental computational units of cognition - integrating Event Segmentation theory with metastability framework (arXiv:2605.31473v1, May 2026).
Langue du texte source : anglais
Metastable neural states as computational units of cognition methodology. Synthesizes event segmentation theory with metastable neural activity, revealing spatio-temporally nested hierarchies and predictive model-driven state transitions. Use when: metastable…
Langue du texte source : anglais
利用MICrONS功能连接组学数据构建生物合理的RNN。整合皮质几何、解剖连接和功能关系作为归纳偏置，实现更有效的学习和收敛到生物计算的组织原则。
Langue du texte source : chinois
Joint latent clustering anomaly detection for multimodal cyber-physical systems (CPS). Models normal behaviour under the MIIM assumption set (Massive, Implicit, Imbalanced Multimodality) with explicit Gaussian-mixture mode clustering in latent space, scored…
Langue du texte source : anglais
Mind-Omni unified multi-task framework for Brain-Vision-Language modeling via discrete diffusion
Langue du texte source : anglais
Skill based on the Mind2Drive paper — a framework for predicting driver intentions from EEG signals captured during real-world on-road driving. Covers multi-sensor synchronization, EEG preprocessing for driving contexts, deep learning architecture comparison…
Langue du texte source : anglais
Tri-modal contrastive framework (EEG, vision, language) for zero-shot visual decoding. Achieves 54.1% Top-1 accuracy on 200-way benchmark, massively exceeding prior baselines. Use for: brain-computer interface visual reconstruction, EEG-based image retrieval,…
Langue du texte source : anglais
MINE (Mechanistically Interpretable Neural Encoding) — a framework that applies mechanistic interpretability tools from LLMs to vision encoding models, revealing fine-grained functional selectivity at the voxel level in human visual cortex. (arXiv:2605.16468)
Langue du texte source : anglais
Mechanistically Interpretable Neural Encoding (MINE) — applying mechanistic interpretability tools (feature attribution, counterfactual editing) to open the black box of voxel-level neural encoding models. Use when: (1) analyzing which image features drive…
Langue du texte source : anglais
Interacting branching model of neural network dynamics with hierarchy of analytical mean-field approximations. Characterizes nonequilibrium phase transitions between disorder and ordered phases, exhibits criticality and self-organized dynamics relevant to…
Langue du texte source : anglais
MIRAGE methodology — robust multi-modal architecture for translating fMRI-to-image models from seen visual decoding to mental imagery reconstruction. Demonstrates that SOTA on seen images doesn't guarantee SOTA on mental imagery and proposes a multi-modal,…
Langue du texte source : anglais
MIRAGE methodology — robust multi-modal architecture for translating fMRI-to-image models from vision decoding to mental image reconstruction. Uses linear backbone + multi-modal text/image features with diffusion model; achieves SOTA on NSD-Imagery benchmark.…
Langue du texte source : anglais
MIRAGE - Adaptive multimodal gating framework for whole-brain fMRI encoding. Integrates visual, auditory, and linguistic information via native multimodal backbone with layer-wise feature gating. Predicts brain responses to naturalistic audiovisual stimuli…
Langue du texte source : anglais
ML-guided Clifford noise reduction for Hamiltonian simulations using mid-circuit measurements. Use when optimizing quantum circuit noise, designing stabilizer verification protocols, reducing logical error rates in encoded quantum operations, or applying ML…
Langue du texte source : anglais
Machine Learning Methods for Studying Latent Neural Activity Dynamics - IJCAI 2026 survey综述机器学习研究神经种群潜伏动力学结构的方法论，涵盖单区域潜伏动力学（LDS/RNN/Neural ODE）、多区域通信、行为对齐建模、神经基础模型（Transformer/扩散模型）
Langue du texte source : chinois
Machine learning methodology for finding optimal quantum error correction (QEC) thresholds. Combines ML search strategies with quantum error correction analysis to determine noise thresholds where QEC codes break down.
Langue du texte source : anglais
Machine Learning approaches for Quantum Error Correction (QEC). Use when researching, designing, or implementing ML-assisted QEC systems including: (1) diffusion models for error decoding (DiffQEC pattern), (2) reinforcement learning for QEC control and…
Langue du texte source : anglais
Maximum Likelihood Decoding of QEC codes — unified survey via statistical mechanics, tensor networks, and AI
Langue du texte source : anglais
MLE-Toolbox: Comprehensive open-source MATLAB toolbox for end-to-end EEG/MEG analysis with source localization, connectivity analysis, and ML classifiers. Activation: MLE-Toolbox, EEG analysis, MEG analysis, source localization, brain network analysis,…
Langue du texte source : anglais
Task-conditioned probing methodology for evaluating brain alignment of instruction-tuned multimodal LLMs (MLLMs) using fMRI. Activation: brain-MLLM alignment, instruction-tuned MLLM, task-conditioned probing, brain encoding, fMRI-MLLM, voxel-wise encoding
Langue du texte source : anglais
Meta-Cognitive Memory Policy Optimization (MMPO) for long-horizon LLM agents using Belief Entropy as self-supervised proxy.
Langue du texte source : anglais
MoDAl (Modality Decorrelation and Alignment) framework for self-supervised neural modality discovery in speech neuroprosthesis. Uses contrastive alignment with LLM text embeddings + decorrelation loss to discover complementary neurolinguistic modalities from…
Langue du texte source : anglais