Machine Learning最新文献

{"title":"Your Next State-of-the-Art Could Come from Another Domain: A Cross-Domain Analysis of Hierarchical Text Classification.","authors":"Nan Li, Bo Kang, Tijl De Bie","doi":"10.1007/s10994-025-06993-w","DOIUrl":"https://doi.org/10.1007/s10994-025-06993-w","url":null,"abstract":"<p><p>Text classification with hierarchical labels is a prevalent and challenging task in natural language processing. Examples include assigning ICD codes to patient records, tagging patents with IPC classes, assigning EUROVOC descriptors to European legal texts, and more. Despite the prevalence of hierarchical text classification problems, a comprehensive understanding of state-of-the-art methods across different application domains has been lacking. In this paper, we propose a unified methodology to break down the boundaries between these different domains, thus enabling cross-domain transfer of innovative ideas. We first construct a <i>Unified Framework</i> that translates distinct domain-specific methods into a common architectural language. Applying this framework, we conduct a comprehensive <i>Cross-Domain Benchmark</i> that exposes architectural gaps often overlooked in single-domain studies. We then demonstrate the framework's practical utility through a validation case study, where we synthesize a new state-of-the-art hierarchical text classification method by combining submodules that were developed for the medical and legal domains. Our extensive empirical analysis yields key insights and guidelines, confirming the necessity of cross-domain learning for designing effective methods. Our code and datasets are publicly available at https://github.com/aida-ugent/cross-domain-HTC.</p>","PeriodicalId":49900,"journal":{"name":"Machine Learning","volume":"115 4","pages":"80"},"PeriodicalIF":2.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13038460/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147610447","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Linear Causal Discovery with Interventional Constraints.","authors":"Zhigao Guo, Feng Dong","doi":"10.1007/s10994-026-06998-z","DOIUrl":"https://doi.org/10.1007/s10994-026-06998-z","url":null,"abstract":"<p><p>Incorporating causal knowledge and mechanisms is essential for refining causal models and improving downstream tasks, such as designing new treatments. In this paper, we introduce a novel concept in causal discovery, termed <i>interventional constraints</i>, which differs fundamentally from interventional data. While interventional data require direct perturbations of variables, interventional constraints encode high-level causal knowledge in the form of inequality constraints on causal effects. For instance, in the Sachs dataset, Akt has been shown to be <i>activated</i> by PIP3, meaning PIP3 exerts a <i>positive</i> causal effect on Akt. Existing causal discovery methods allow enforcing structural constraints (e.g., requiring a causal path from PIP3 to Akt), but they may still produce incorrect causal conclusions, such as learning that \"PIP3 <i>inhibits</i> Akt.\" Interventional constraints bridge this gap by explicitly constraining the total causal effect between variable pairs, ensuring learned models respect known causal influences. To formalize interventional constraints, we adopt a metric to quantify total causal effects for <i>linear</i> causal models and formulate the problem as a constrained optimization task, solved using a two-stage constrained optimization method. We evaluate our approach on real-world datasets and demonstrate that integrating interventional constraints not only improves model accuracy and ensures consistency with established findings, making models more explainable, but also facilitates the discovery of new causal relationships that would otherwise be costly to identify.</p>","PeriodicalId":49900,"journal":{"name":"Machine Learning","volume":"115 3","pages":"35"},"PeriodicalIF":2.9,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12913306/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146229542","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep latent force models: ODE-based process convolutions for Bayesian deep learning.","authors":"Thomas Baldwin-McDonald, Xinxing Shi, Mingxin Shen, Mauricio A Álvarez","doi":"10.1007/s10994-025-06824-y","DOIUrl":"10.1007/s10994-025-06824-y","url":null,"abstract":"<p><p>Modelling the behaviour of highly nonlinear dynamical systems with robust uncertainty quantification is a challenging task which typically requires approaches specifically designed to address the problem at hand. We introduce a domain-agnostic model to address this issue termed the deep latent force model (DLFM), a deep Gaussian process with physics-informed kernels at each layer, derived from ordinary differential equations using the framework of process convolutions. Two distinct formulations of the DLFM are presented which utilise weight-space and variational inducing points-based Gaussian process approximations, both of which are amenable to doubly stochastic variational inference. We present empirical evidence of the capability of the DLFM to capture the dynamics present in highly nonlinear real-world multi-output time series data. Additionally, we find that the DLFM is capable of achieving comparable performance to a range of non-physics-informed probabilistic models on benchmark univariate regression tasks. We also empirically assess the negative impact of the inducing points framework on the extrapolation capabilities of LFM-based models.</p>","PeriodicalId":49900,"journal":{"name":"Machine Learning","volume":"114 8","pages":"192"},"PeriodicalIF":4.3,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12263784/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144660909","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Ensuring medical AI safety: interpretability-driven detection and mitigation of spurious model behavior and associated data.","authors":"Frederik Pahde, Thomas Wiegand, Sebastian Lapuschkin, Wojciech Samek","doi":"10.1007/s10994-025-06834-w","DOIUrl":"10.1007/s10994-025-06834-w","url":null,"abstract":"<p><p>Deep neural networks are increasingly employed in high-stakes medical applications, despite their tendency for shortcut learning in the presence of spurious correlations, which can have potentially fatal consequences in practice. Whereas a multitude of works address either the detection or mitigation of such shortcut behavior in isolation, the Reveal2Revise approach provides a comprehensive bias mitigation framework combining these steps. However, effectively addressing these biases often requires substantial labeling efforts from domain experts. In this work, we review the steps of the Reveal2Revise framework and enhance it with semi-automated interpretability-based bias annotation capabilities. This includes methods for the sample- and feature-level bias annotation, providing valuable information for bias mitigation methods to unlearn the undesired shortcut behavior. We show the applicability of the framework using four medical datasets across two modalities, featuring controlled and real-world spurious correlations caused by data artifacts. We successfully identify and mitigate these biases in VGG16, ResNet50, and contemporary Vision Transformer models, ultimately increasing their robustness and applicability for real-world medical tasks. Our code is available at https://github.com/frederikpahde/medical-ai-safety.</p>","PeriodicalId":49900,"journal":{"name":"Machine Learning","volume":"114 9","pages":"206"},"PeriodicalIF":2.9,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12343733/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144856810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mining exceptional social behavior on attributed interaction networks.","authors":"Martin Atzmueller, Carolina Centeio Jorge, Cláudio Rebelo de Sá, Behzad M Heravi, Jenny L Gibson, Rosaldo J F Rossetti","doi":"10.1007/s10994-025-06831-z","DOIUrl":"10.1007/s10994-025-06831-z","url":null,"abstract":"<p><p>Social interactions are prevalent in our lives. These can be observed, e. g., online using social media, however, also offline specifically using sensors. In such contexts, typically time-stamped interactions are recorded, which can also be inferred from real-time location of humans. Such interaction data can then be modeled as so-called social interaction networks. For their analysis, a variety of different approaches can be applied. A prominent research direction is then the detection of patterns describing specific subgroups with exceptional behavioral characteristics, given some measure of interest. In the standard case of plain graphs modeling the interaction networks, methods for identifying such subgroups mainly focus on structural characteristics of the network and/or the induced subgraph. For attributed social networks, then additional attributive information can be exploited. This paper proposes to focus on the dyadic structure of the attributed social interaction networks, thus enabling a compositional perspective for identifying interesting subgroup patterns. Specifically, we can then analyze spatio-temporal data modeled as attributed social interaction networks for identifying exceptional social behavior. The presented approach adapts local pattern mining using subgroup discovery to the dyadic setting, exploiting attribute information of the spatio-temporal attributed interaction networks. With this, specific characteristics of social interactions are considered, i. e., duration and frequency, for identifying subgroups capturing social behavior that deviates from the norm. For subgroup discovery, we propose according interestingness measures in the form of seven novel quality functions and discuss their properties. In our experimentation, we perform an evaluation demonstrating the efficacy of the presented approach using four real-world datasets on face-to-face interactions in academic conferencing as well as school playground contexts. Our results indicate that the proposed method returns interesting, meaningful, and valid findings and results.</p>","PeriodicalId":49900,"journal":{"name":"Machine Learning","volume":"114 11","pages":"243"},"PeriodicalIF":2.9,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12513876/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145281580","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Boolean matrix logic programming for active learning of gene functions in genome-scale metabolic network models.","authors":"Lun Ai, Stephen H Muggleton, Shi-Shun Liang, Geoff S Baldwin","doi":"10.1007/s10994-025-06868-0","DOIUrl":"10.1007/s10994-025-06868-0","url":null,"abstract":"<p><p>Reasoning about hypotheses and updating knowledge through empirical observations are central to scientific discovery. In this work, we applied logic-based machine learning methods to drive biological discovery by guiding experimentation. Genome-scale metabolic network models (GEMs) - comprehensive representations of metabolic genes and reactions - are widely used to evaluate genetic engineering of biological systems. However, GEMs often fail to accurately predict the behaviour of genetically engineered cells, primarily due to incomplete annotations of gene interactions. The task of learning the intricate genetic interactions within GEMs presents computational and empirical challenges. To efficiently predict using GEM, we describe a novel approach called Boolean Matrix Logic Programming (BMLP) by leveraging Boolean matrices to evaluate large logic programs. We developed a new system, [Formula: see text], which guides cost-effective experimentation and uses interpretable logic programs to encode a state-of-the-art GEM of a model bacterial organism. Notably, [Formula: see text] successfully learned the interaction between a gene pair with fewer training examples than random experimentation, overcoming the increase in experimental design space. [Formula: see text] enables rapid optimisation of metabolic models to reliably engineer biological systems for producing useful compounds. It offers a realistic approach to creating a self-driving lab for biological discovery, which would then facilitate microbial engineering for practical applications.</p>","PeriodicalId":49900,"journal":{"name":"Machine Learning","volume":"114 11","pages":"254"},"PeriodicalIF":2.9,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12535945/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145349575","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Computing the distance between unbalanced distributions: the flat metric.","authors":"Henri Schmidt, Christian Düll","doi":"10.1007/s10994-025-06828-8","DOIUrl":"10.1007/s10994-025-06828-8","url":null,"abstract":"<p><p>We provide an implementation to compute the flat metric in any dimension. The flat metric, also called dual bounded Lipschitz distance, generalizes the well-known Wasserstein distance <math><msub><mi>W</mi> <mn>1</mn></msub> </math> to the case that the distributions are of unequal total mass. Thus, our implementation adapts very well to mass differences and uses them to distinguish between different distributions. This is of particular interest for unbalanced optimal transport tasks and for the analysis of data distributions where the sample size is important or normalization is not possible. The core of the method is based on a neural network to determine an optimal test function realizing the distance between two given measures. Special focus was put on achieving comparability of pairwise computed distances from independently trained networks. We tested the quality of the output in several experiments where ground truth was available as well as with simulated data.</p>","PeriodicalId":49900,"journal":{"name":"Machine Learning","volume":"114 8","pages":"195"},"PeriodicalIF":2.9,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12289810/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144734905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Interpretable optimisation-based approach for hyper-box classification.","authors":"Georgios I Liapis, Sophia Tsoka, Lazaros G Papageorgiou","doi":"10.1007/s10994-024-06643-7","DOIUrl":"10.1007/s10994-024-06643-7","url":null,"abstract":"<p><p>Data classification is considered a fundamental research subject within the machine learning community. Researchers seek the improvement of machine learning algorithms in not only accuracy, but also interpretability. Interpretable algorithms allow humans to easily understand the decisions that a machine learning model makes, which is challenging for black box models. Mathematical programming-based classification algorithms have attracted considerable attention due to their ability to effectively compete with leading-edge algorithms in terms of both accuracy and interpretability. Meanwhile, the training of a hyper-box classifier can be mathematically formulated as a Mixed Integer Linear Programming (MILP) model and the predictions combine accuracy and interpretability. In this work, an optimisation-based approach is proposed for multi-class data classification using a hyper-box representation, thus facilitating the extraction of compact IF-THEN rules. The key novelty of our approach lies in the minimisation of the number and length of the generated rules for enhanced interpretability. Through a number of real-world datasets, it is demonstrated that the algorithm exhibits favorable performance when compared to well-known alternatives in terms of prediction accuracy and rule set simplicity.</p>","PeriodicalId":49900,"journal":{"name":"Machine Learning","volume":"114 3","pages":"51"},"PeriodicalIF":4.3,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11861270/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143525101","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Offline reinforcement learning for learning to dispatch for job shop scheduling.","authors":"Jesse van Remmerden, Zaharah Bukhsh, Yingqian Zhang","doi":"10.1007/s10994-025-06826-w","DOIUrl":"10.1007/s10994-025-06826-w","url":null,"abstract":"<p><p>The Job Shop Scheduling Problem (JSSP) is a complex combinatorial optimization problem. While online Reinforcement Learning (RL) has shown promise by quickly finding acceptable solutions for JSSP, it faces key limitations: it requires extensive training interactions from scratch leading to sample inefficiency, cannot leverage existing high-quality solutions from traditional methods like Constraint Programming (CP), and require simulated environments to train in, which are impracticable to build for complex scheduling environments. We introduce Offline Learned Dispatching (Offline-LD), an offline reinforcement learning approach for JSSP, which addresses these limitations by learning from historical scheduling data. Our approach is motivated by scenarios where historical scheduling data and expert solutions are available or scenarios where online training of RL approaches with simulated environments is impracticable. Offline-LD introduces maskable variants of two Q-learning methods, namely, Maskable Quantile Regression DQN (mQRDQN) and discrete maskable Soft Actor-Critic (d-mSAC), that are able to learn from historical data, through Conservative Q-Learning (CQL), whereby we present a novel entropy bonus modification for d-mSAC, for maskable action spaces. Moreover, we introduce a novel reward normalization method for JSSP in an offline RL setting. Our experiments demonstrate that Offline-LD outperforms online RL on both generated and benchmark instances when trained on only 100 solutions generated by CP. Notably, introducing noise to the expert dataset yields comparable or superior results to using the expert dataset, with the same amount of instances, a promising finding for real-world applications, where data is inherently noisy and imperfect.</p>","PeriodicalId":49900,"journal":{"name":"Machine Learning","volume":"114 8","pages":"191"},"PeriodicalIF":4.3,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12263752/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144660910","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"On metafeatures’ ability of implicit concept identification","authors":"Joanna Komorniczak, Paweł Ksieniewicz","doi":"10.1007/s10994-024-06612-0","DOIUrl":"https://doi.org/10.1007/s10994-024-06612-0","url":null,"abstract":"<p>Concept drift in data stream processing remains an intriguing challenge and states a popular research topic. Methods that actively process data streams usually employ drift detectors, whose performance is often based on monitoring the variability of different stream properties. This publication provides an overview and analysis of metafeatures variability describing data streams with concept drifts. Five experiments conducted on synthetic, semi-synthetic, and real-world data streams examine the ability of over 160 metafeatures from 9 categories to recognize concepts in non-stationary data streams. The work reveals the distinctions in the considered sources of streams and specifies 17 metafeatures with a high ability of concept identification.</p>","PeriodicalId":49900,"journal":{"name":"Machine Learning","volume":"51 1","pages":""},"PeriodicalIF":7.5,"publicationDate":"2024-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142253817","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}