![[2602.15229] tensorFM: Low-Rank Approximations of Cross-Order Feature Interactions](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.15229] tensorFM: Low-Rank Approximations of Cross-Order Feature Interactions
Summary
The paper introduces tensorFM, a model designed for efficient low-rank approximations of cross-order feature interactions in tabular categorical data, demonstrating competitive performance in prediction tasks.
Why It Matters
As prediction problems involving categorical data are prevalent in fields like online advertising and social sciences, tensorFM offers a novel approach that enhances performance while maintaining low latency, making it relevant for real-time applications.
Key Takeaways
- tensorFM captures high-order interactions between categorical attributes effectively.
- The model generalizes field-weighted factorization machines, enhancing their capabilities.
- Empirical results show tensorFM's competitive performance against state-of-the-art methods.
- Its low latency makes it suitable for time-sensitive applications like online advertising.
- The model addresses common challenges in prediction tasks involving tabular categorical data.
Computer Science > Machine Learning arXiv:2602.15229 (cs) [Submitted on 16 Feb 2026] Title:tensorFM: Low-Rank Approximations of Cross-Order Feature Interactions Authors:Alessio Mazzetto (1), Mohammad Mahdi Khalili (2 and 3), Laura Fee Nern (3), Michael Viderman (3), Alex Shtoff (4), Krzysztof Dembczyński (3 and 5) ((1) Brown University, (2) Ohio State University, (3) Yahoo Research, (4) Technology Innovation Institute, (5) Poznan University of Technology) View a PDF of the paper titled tensorFM: Low-Rank Approximations of Cross-Order Feature Interactions, by Alessio Mazzetto (1) and 9 other authors View PDF HTML (experimental) Abstract:We address prediction problems on tabular categorical data, where each instance is defined by multiple categorical attributes, each taking values from a finite set. These attributes are often referred to as fields, and their categorical values as features. Such problems frequently arise in practical applications, including click-through rate prediction and social sciences. We introduce and analyze {tensorFM}, a new model that efficiently captures high-order interactions between attributes via a low-rank tensor approximation representing the strength of these interactions. Our model generalizes field-weighted factorization machines. Empirically, tensorFM demonstrates competitive performance with state-of-the-art methods. Additionally, its low latency makes it well-suited for time-sensitive applications, such as online advertising. Subjects: M...
