Data & Analytics

Neural Cellular Automata (NCA) attractor stability and geometry analysis methodology. Analyzes learned attractor states in NCAs beyond visual similarity, measuring basin of attraction, convergence properties, and geometric structure. Activation: NCA attractor, neural cellular automata stability, attractor basin analysis, NCA convergence.

Excitation-driven data generation and distributed control optimization for building thermal systems and district heating networks. Combines BuilDyn framework (arXiv:2605.29849) and distributed NMPC with ADMM (arXiv:2605.29841).

Distributionally Robust Data-Driven Predictive Control (DR-DDPC) methodology for stochastic LTI systems with unknown dynamics and disturbance distributions. Combines subspace predictive control (SPC) with distributionally robust optimization using Wasserstein ambiguity sets. Use when: designing robust controllers under uncertainty, implementing data-driven MPC, handling stochastic disturbances with unknown distributions, or applying distributionally robust optimization to control systems.

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 optimal control, stability analysis, surrogate model control. Activation: discounted MPC, plant-model mismatch, robust MPC, infinite-horizon control, suboptimality analysis.

Comprehensive framework for anomaly detection in Cyber-Physical Systems (CPS) security. Covers model-based, data-driven, statistical, and hybrid approaches for detecting cyber threats in critical infrastructure. Activation: CPS security, anomaly detection, cyber-physical systems, intrusion detection, industrial control system security.

Convex Hybrid Modeling methodology using operator theory for process control and systems engineering. Formulates convex learning problems that combine model interpretability with system identification efficiency. Covers three settings: (1) regularization around a reference model, (2) restriction on interpretable subspaces, (3) kernel-based mixture models on interpretable manifolds. Use when: building interpretable control models, combining physics-based and data-driven modeling, designing hyb...

BCMI-driven motion control detection using EEG-based machine learning and interaction entropy for high-order brain networks during music-assisted driving

ARFIMA decomposition of stride-to-stride fluctuations in human walking for sensorimotor control analysis. Activation triggers: stride fluctuations, human gait, DFA, ARFIMA, fractal analysis, sensorimotor control

Proposes that active sensing (energy expenditure for information) is not driven by sensory goals but is necessary for task-level control. Integrates empirical data and control theory to explain explore-exploit mode switching in biological sensorimotor systems. Use when researching active sensing, sensorimotor control, control theory in neuroscience, explore-exploit tradeoffs, or bio-inspired robotics.

Complex dynamics in the Sherrington-Kirkpatrick (SK) game methodology — game-theoretic foundation for adaptive learning in disordered many-player systems with random payoff matrices. Generalizes the SK spin-glass model to game theory with random-field bias, grand-canonical abstention, and convergence/volatility phase diagram. Bridges spin-glass neural network theory, reinforcement learning, and game theory. arXiv:2607.02422

Sharma-Mittal entropy framework bridging information theory, black hole thermodynamics, and infrared gravity modifications. Derives modified gravitational force laws from generalized entropy, reproduces MOND-like regime. Activates: sharma-mittal entropy, generalized entropy, emergent gravity, MOND, black hole thermodynamics, information bounds, infrared gravity, entropic gravity

Semiclassical methods connecting quantum statistical mechanics to analytic number theory. Uses trace formula and periodic orbit theory to study integer partitions. Activation: semiclassical, integer partitions, density of states, number theory, periodic orbit, trace formula, Pythagorean triples.

SABER framework for semantic-aligned brain network analysis via multi-scale hypergraphs. Actively integrates LLM-derived semantics into brain network prediction, combining global self-attention, multi-scale hypergraph construction, and decision-level semantic alignment for improved brain disease diagnosis. Use when building brain network classifiers, fMRI/EEG analysis pipelines, or LLM-brain integration systems.

Finite combinatorial analogue of Gödel's incompleteness theorems within Boolean K-SAT. Proves self-referential hardness exhibits physical invariance precluding quantum shortcuts due to necessity of global semantic analysis, and delineates scaling bottleneck for ML on lossy local compression.

Formalizes how attractor networks emerge from the free energy principle applied to universal partitioning of random dynamical systems. Results in self-orthogonalizing attractor representations, biologically plausible multi-level Bayesian active inference. Use when: studying attractor dynamics in neural networks, free energy principle applications, Bayesian active inference models, biologically plausible learning, Boltzmann Machine variants, self-organizing neural dynamics. Triggered by: free ...

Machine learning models are increasingly deployed in high-stakes domains, raising concerns about both privacy and fairness. Differential Privacy (DP) has become a gold standard for privacy-preserving. Based on arXiv:2607.07471.

**arXiv ID:** 2212.06735 **Authors:** Andrea Falanti, Eugenio Lomurno, Danilo Ardagna, Matteo Matteucci **Published:** 2022-12-13T17:14:14Z **Abstract:** The automated machine learning (AutoML) field has become increasingly relevant in recent years. These algorithms can develop models without the need for expert knowledge, facilitating the application of machine learning techniques in the industry. Neural Architecture Search (NAS) exploits deep learning techniques to autonomously produce neur...

**arXiv ID:** 2112.12251 **Authors:** Amin Banitalebi-Dehkordi, Yong Zhang **Published:** 2021-12-22T22:40:13Z **Abstract:** The 2021 NeurIPS Machine Learning for Combinatorial Optimization (ML4CO) competition was designed with the goal of improving state-of-the-art combinatorial optimization solvers by replacing key heuristic components with machine learning models. The competition's main scientific question was the following: is machine learning a viable option for improving traditional com...

Federated learning has become a popular paradigm for privacy protection and edge-based machine learning. However, defending against differential attac... 触发词: 联邦学习, 控制系统.

**arXiv ID:** 2505.04468 **Authors:** Hyeju Shin, Vincent-Daniel, Kyudan Jung, Seongwon Yun **Published:** 2025-05-07T14:38:58Z **Abstract:** Differential Privacy (DP) has emerged as a key framework for protecting sensitive data in machine learning, but standard DP-SGD often suffers from significant accuracy loss due to injected noise. To address this limitation, we introduce the FFT-Enhanced Kalman Filter (FFTKF), a differentially private optimization method that improves gradient quality w...

**arXiv ID:** 2101.02442 **Authors:** Clément Leroy, Eric Anquetil, Nathalie Girard **Published:** 2021-01-07T09:21:27Z **Abstract:** Working with a non-stationary stream of data requires for the analysis system to evolve its model (the parameters as well as the structure) over time. In particular, concept drifts can occur, which makes it necessary to forget knowledge that has become obsolete. However, the forgetting is subjected to the stability-plasticity dilemma, that is, increasing forget...

Derived from arXiv:2607.17897 - Distributional Soft Bellman Operator under the Cramér Geometry

**arXiv ID:** 1607.01719 **Authors:** Baochen Sun, Kate Saenko **Published:** 2016-07-06T17:35:55Z **Abstract:** Deep neural networks are able to learn powerful representations from large quantities of labeled input data, however they cannot always generalize well across changes in input distributions. Domain adaptation algorithms have been proposed to compensate for the degradation in performance due to domain shift. In this paper, we address the case when the target domain is unlabeled, req...

**arXiv ID:** 2212.09030 **Authors:** Slawek Smyl, Grzegorz Dudek, Paweł Pełka **Published:** 2022-12-18T07:42:48Z **Abstract:** In this paper, we propose a new short-term load forecasting (STLF) model based on contextually enhanced hybrid and hierarchical architecture combining exponential smoothing (ES) and a recurrent neural network (RNN). The model is composed of two simultaneously trained tracks: the context track and the main track. The context track introduces additional information to...