The transition from traditional power grids to smart grids, significant increase in the use of renewable energy sources, and soaring electricity prices has triggered a digital transformation of the energy infrastructure that enables new, data driven, applications often supported by machine learning models. However, the majority of the developed machine learning models rely on univariate data. To date, a structured study considering the role meta-data and additional measurements resulting in multivariate data is missing. In this paper we propose a taxonomy that identifies and structures various types of data related to energy applications. The taxonomy can be used to guide application specific data model development for training machine learning models. Focusing on a household electricity forecasting application, we validate the e ff ectiveness of the proposed taxonomy in guiding the selection of the features for various types of models. Finally, using a feature importance techniques, we explain individual feature contributions to the forecasting accuracy.1. Introduction The transition from traditional power grids to smart grids, significant increase in the use of renewable energy sources, and soaring electricity prices has led to an increase in complexity [1], particularly with the adoption of smart meters (SMs), energy management systems (EMSes), and intelligent electronic devices (IEDs) at the low voltage (L V) level. These devices enable innovative energy [2] and non-energy applications [3, 4], such as energy cost optimization and matching consumption with self-production from renewable energy sources. On the distribution system operator (DSO) side of the L V grid, reliability and latency are the main challenges, and complete ob-servability of the L V grid for each substation is crucial.

This paper explores second-order optimization methods in Federated Learning (FL), addressing the critical challenges of slow convergence and the excessive communication rounds required to achieve optimal performance from the global model. While existing surveys in FL primarily focus on challenges related to statistical and device label heterogeneity, as well as privacy and security concerns in first-order FL methods, less attention has been given to the issue of slow model training. This slow training often leads to the need for excessive communication rounds or increased communication costs, particularly when data across clients are highly heterogeneous. In this paper, we examine various FL methods that leverage second-order optimization to accelerate the training process. We provide a comprehensive categorization of state-of-the-art second-order FL methods and compare their performance based on convergence speed, computational cost, memory usage, transmission overhead, and generalization of the global model. Our findings show the potential of incorporating Hessian curvature through second-order optimization into FL and highlight key challenges, such as the efficient utilization of Hessian and its inverse in FL. This work lays the groundwork for future research aimed at developing scalable and efficient federated optimization methods for improving the training of the global model in FL.

We introduce dro, an open-source Python library for distributionally robust optimization (DRO) for regression and classification problems. The library implements 14 DRO formulations and 9 backbone models, enabling 79 distinct DRO methods. Furthermore, dro is compatible with both scikit-learn and PyTorch. Through vectorization and optimization approximation techniques, dro reduces runtime by 10x to over 1000x compared to baseline implementations on large-scale datasets. Comprehensive documentation is available at https://python-dro.org.

Linear regression is a foundational algorithm in statistics and machine learning, widely employed for modeling the linear relationship between a dependent (or target) variable and one or more independent (or explanatory) variables (e.g., Montgomery et al. [15] and Weisberg [23]). Its simplicity, interpretability, and efficiency have made it an indispensable tool across diverse fields such as economics, engineering, biology, and social sciences. The core objective in linear regression is to determine the optimal set of parameters (or weights) for the independent variables that best predict the dependent variable, typically by minimizing the sum of squared differences between observed and predicted values--a criterion known as Ordinary Least Squares (OLS). Figure 1 provides a conceptual illustration of linear regression. Once the problem is formulated, the next crucial step is to solve for these optimal parameters. Two predominant strategies for achieving this are: The Moore-Penrose pseudoinverse, which provides a direct, analytical solution.

We present a thermodynamic interpretation of the stationary behavior of stochastic gradient descent (SGD) under fixed learning rates (LRs) in neural network training. We show that SGD implicitly minimizes a free energy function $F=U-TS$, balancing training loss $U$ and the entropy of the weights distribution $S$, with temperature $T$ determined by the LR. This perspective offers a new lens on why high LRs prevent training from converging to the loss minima and how different LRs lead to stabilization at different loss levels. We empirically validate the free energy framework on both underparameterized (UP) and overparameterized (OP) models. UP models consistently follow free energy minimization, with temperature increasing monotonically with LR, while for OP models, the temperature effectively drops to zero at low LRs, causing SGD to minimize the loss directly and converge to an optimum. We attribute this mismatch to differences in the signal-to-noise ratio of stochastic gradients near optima, supported by both a toy example and neural network experiments.

Fast and scalable alignment of time series is a fundamental challenge in many domains. The standard solution, Dynamic Time Warping (DTW), struggles with poor scalability and sensitivity to noise. We introduce TimePoint, a self-supervised method that dramatically accelerates DTW-based alignment while typically improving alignment accuracy by learning keypoints and descriptors from synthetic data. Inspired by 2D keypoint detection but carefully adapted to the unique challenges of 1D signals, TimePoint leverages efficient 1D diffeomorphisms, which effectively model nonlinear time warping, to generate realistic training data. This approach, along with fully convolutional and wavelet convolutional architectures, enables the extraction of informative keypoints and descriptors. Applying DTW to these sparse representations yield major speedups and typically higher alignment accuracy than standard DTW applied to the full signals. TimePoint demonstrates strong generalization to real-world time series when trained solely on synthetic data, and further improves with fine-tuning on real data. Extensive experiments demonstrate that TimePoint consistently achieves faster and more accurate alignments than standard DTW, making it a scalable solution for time-series analysis. Our code is available at https://github.com/BGU-CS-VIL/TimePoint

Abstract--Depression is a debilitating mood disorder negatively impacting millions worldwide. While researchers have explored multiple verbal and non-verbal behavioural cues for automated depression assessment, head motion has received little attention thus far. Further, the common practice of validating machine learning models via a single dataset can limit model generalisability . This work examines the effectiveness and generalisability of models utilising elementary head motion units, termed kinemes, for depression severity estimation. Specifically, we consider three depression datasets from different western cultures (German: AVEC2013, Australian: Blackdog and American: Pitt datasets) with varied contextual and recording settings to investigate the generalisability of the derived kineme patterns via two methods: (i) k-fold cross-validation over individual/multiple datasets, and (ii) model reuse on other datasets. Evaluating classification and regression performance with classical machine learning methods, our results show that: (1) head motion patterns are efficient biomarkers for estimating depression severity, achieving highly competitive performance for both classification and regression tasks on a variety of datasets, including achieving the second best Mean Absolute Error (MAE) on the AVEC2013 dataset, and (2) kineme-based features are more generalisable than (a) raw head motion descriptors for binary severity classification, and (b) other visual behavioural cues for severity estimation (regression).

Quantum sensing exploits non-classical effects to overcome limitations of classical sensors, with applications ranging from gravitational-wave detection to nanoscale imaging. However, practical quantum sensors built on noisy intermediate-scale quantum (NISQ) devices face significant noise and sampling constraints, and current variational quantum sensing (VQS) methods lack rigorous performance guarantees. This paper proposes an online control framework for VQS that dynamically updates the variational parameters while providing deterministic error bars on the estimates. By leveraging online conformal inference techniques, the approach produces sequential estimation sets with a guaranteed long-term risk level. Experiments on a quantum magnetometry task confirm that the proposed dynamic VQS approach maintains the required reliability over time, while still yielding precise estimates. The results demonstrate the practical benefits of combining variational quantum algorithms with online conformal inference to achieve reliable quantum sensing on NISQ devices.

Determining conditional independence (CI) relationships between random variables is a fundamental yet challenging task in machine learning and statistics, especially in high-dimensional settings. Existing generative model-based CI testing methods, such as those utilizing generative adversarial networks (GANs), often struggle with undesirable modeling of conditional distributions and training instability, resulting in subpar performance. To address these issues, we propose a novel CI testing method via score-based generative modeling, which achieves precise Type I error control and strong testing power. Concretely, we first employ a sliced conditional score matching scheme to accurately estimate conditional score and use Langevin dynamics conditional sampling to generate null hypothesis samples, ensuring precise Type I error control. Then, we incorporate a goodness-of-fit stage into the method to verify generated samples and enhance interpretability in practice. We theoretically establish the error bound of conditional distributions modeled by score-based generative models and prove the validity of our CI tests. Extensive experiments on both synthetic and real-world datasets show that our method significantly outperforms existing state-of-the-art methods, providing a promising way to revitalize generative model-based CI testing.

Generalized Category Discovery (GCD) aims to classify both known and novel categories using partially labeled data that contains only known classes. Despite achieving strong performance on existing benchmarks, current textual GCD methods lack sufficient validation in realistic settings. We introduce Event-Centric GCD (EC-GCD), characterized by long, complex narratives and highly imbalanced class distributions, posing two main challenges: (1) divergent clustering versus classification groupings caused by subjective criteria, and (2) Unfair alignment for minority classes. To tackle these, we propose PaMA, a framework leveraging LLMs to extract and refine event patterns for improved cluster-class alignment. Additionally, a ranking-filtering-mining pipeline ensures balanced representation of prototypes across imbalanced categories. Evaluations on two EC-GCD benchmarks, including a newly constructed Scam Report dataset, demonstrate that PaMA outperforms prior methods with up to 12.58% H-score gains, while maintaining strong generalization on base GCD datasets.