Aug 2026· Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.2· 0 citations· 1 references
Abstract
Structured data drives enterprise decision-making, yet building predictive pipelines for time-series and tabular modalities requires intensive feature engineering, model selection, and other ''tricks of the trade''. AutoGluon, an open source AutoML system, automates this through multi-layer stack ensembling, providing a unified API that reliably achieves state-of-the-art accuracy across time series and tabular data. Concurrently, Foundation Models for Structured Data (FM4SD) have emerged to push the boundaries of predictive performance, enabling both powerful zero-shot inference and efficient fine-tuning. Specifically, Chronos-2 delivers state-of-the-art multivariate forecasting for time series, while a rapidly growing ecosystem of tabular foundation models (e.g., TabPFN, Mitra, TabICL) transforms how practitioners execute classification and regression tasks. This system-focused tutorial teaches attendees how to operationalize FM4SDs using AutoML across temporal and tabular modalities, with an additional highlight on the newly released Chronos-2. We guide participants through a technical progression: establishing AutoGluon as the orchestration framework, detailing the internal mechanics of Chronos-2 and various tabular FMs, and unifying them into production-ready pipelines.
Large language models are increasingly used as open-ended search operators in evolutionary optimization. We introduce Evolutionary Feature Engineering (EFE), a framework for using LLM-based evolution to discover preprocessing transformations for structured data. EFE represents transformations as Python programs with a standardized fit/transform interface, allowing them to be inserted directly into existing machine learning pipelines. During evolution, candidate programs are refined using dataset context, summary statistics, and downstream performance feedback on validation set. We instantiate EFE in two settings. For time-series forecasting, EFE-Time learns invertible, dataset-specific normalizations that improve off-the-shelf time-series foundation models. It reduces forecasting errors (MASE, WQL, MAE) 3% or more when averaged across datasets and improvements are as much as 19% on the COVID-Deaths dataset. Notably, these improvements occur with recent TSFMs such as Chronos-2. For tabular prediction, EFE-Tab evolves compact feature programs that add useful interpretable features and remove redundant ones, improving or matching existing LLM-based feature-engineering methods. We found EFE-Tab to be particularly effective on classical decision trees, where small sets of evolved features yield competitive accuracy while preserving interpretability. Overall, EFE demonstrates that LLM-based evolution can improve both accuracy and interpretability when automatically tackling structured data.
Ege Onur Taga, Yilin Zhuang, M. E. Ildiz et al.· 0 citations
Developing interpretable models on structured tabular data is difficult because there is a sequence dependency among data pre-processing, model learning, hyperparameter tuning, and result interpretation. The existing AutoML systems focus more on predictive accuracy rather than having an integrated approach towards explainability for different types of learning tasks. In this research, we introduce an Explainable AutoML Pipeline for performing automated data pre-processing, multi-model learning, hyperparameter optimization using Optuna, model selection using a leaderboard, and finally interpreting the results using SHAP. The framework automatically performs preprocessing on the input data, optimizes various candidate models, and determines the most suitable learning algorithm without any human intervention while also offering explanations about the decision-making process on a per-feature level. Model validation is then done by applying it to the California Housing and Titanic datasets. In particular, the optimized LightGBM model scored 0.8470 on the regression R2 metric, while the optimized XGBoost classifier managed to reach 81.56% accuracy, an F1-score of 0.7481, and an AUC of 0.8117. The SHAP analysis successfully detected the most influential predictive features, which increases the model interpretability
A. S, R. A, Shanmathi K. et al.· Journal of ISMAC· 0 citations
Tabular foundation models (TFMs) have moved tabular machine learning from per-dataset training towards amortised in-context inference, fitting a small-to-medium table in a single forward pass without a training run. The 2024–2026 release train, the TabPFN and TabICL lines and challengers such as Mitra, LimiX and Orion, has produced a generation whose architectures, capabilities and limits are documented mainly in preprints, while existing surveys treat these models as a subsection of tabular deep learning or of language-model table understanding. This review is, to our knowledge, the first organised around the current generation. From a corpus of 961 screened records and 98 retained studies, it taxonomises the architectures by pretraining regime, maps the capability space across five axes, isolates the language-model-on-tabular strand for prediction, feature engineering and generation, and summarises openness and deployment. A dedicated critical synthesis then reads the reported capabilities against independent evidence: on the studies reviewed here, tree-based and deep models retain the lead across 142 curated datasets that go beyond the standard independent and identically distributed setting; on 112 datasets, the models attain the highest accuracy but weaker conditional coverage than gradient-boosted trees; and robustness under feature shift, fairness and generation quality remain open. Amortised in-context prediction is thus a working paradigm whose independent evidence has yet to match its benchmark claims.
Sergei O. Kurashkin, V. Tynchenko, Alexey S. Borodulin et al.· Machine Learning and Knowled...· 0 citations
Although Automated Machine Learning (AutoML) systems allow the use of Machine Learning (ML) to automate the design of ML pipelines, they typically search over fixed, task-agnostic configuration spaces, leading to high computational costs. This paper overviews a Ph.D. thesis that proposes a paradigm shift: using Metalearning (MtL) to dynamically build task-specific search spaces. Unlike prior approaches that either optimize within a fixed search space or directly recommend algorithms without an optimization step, this thesis introduces the Dynamic Pipeline CASH problem, which extends the CASH formulation to incorporate meta-model-driven search space creation for pipelines. The thesis contributes a systematic literature review identifying meta-knowledge as the unifying thread across AutoML subfields, applied studies reinforcing the importance of algorithm selection and tuning, a large-scale benchmark of over one million pipeline configurations, and the pymfe package for reproducible meta-feature extraction. These building blocks converge into a novel MtL framework that dynamically reduces search spaces while maintaining competitive performance.
Edesio Alcobaça, A. Carvalho· Anais do XXXIX Concurso de T...· 0 citations
The results show that the LLM's capability dissolves with dimension in a way no noise-corrupted classical learner mimics - which explains why LLMs, so capable elsewhere, keep losing to fifty-year-old baselines on tables, while leaving the mechanism of the prediction as an open question.
Feature engineering plays a key role in determining the performance of machine learning models, but manual feature
design is time-consuming and requires strong domain knowledge. This work presents an automated feature engineering
framework integrated with an interpretable AutoML pipeline, built around the BigFeat methodology. The system automatically
generates new features from existing data using mathematical and logical operators and selects the most stable and relevant
features for learning, while preserving interpretability by maintaining traceable mappings between original and engineered
features. The framework is designed to handle large, high-dimensional datasets with manageable computational overhead.
Automated model selection and hyperparameter tuning across Random Forest, Logistic Regression, and Decision Tree
classifiers are incorporated to optimize predictive performance without manual intervention. The proposed system is intended to
reduce human effort and development time in the feature engineering process while remaining scalable and adaptable to
different datasets. Experimental evaluation on the Madelon dataset, a high-dimensional synthetic benchmark for feature
selection, indicates that the automated and interpretable pipeline performs comparably to, and in some respects favourably
against, baseline feature engineering approaches, demonstrating the practical effectiveness of combining scalable feature
generation with interpretable AutoML.
Ms. CH. Vasavi, Ms. SK. Raqeeba· International Journal for Re...· 0 citations