Research

Paper

AI LLM February 19, 2026

A Theoretical Framework for Modular Learning of Robust Generative Models

Authors

Corinna Cortes, Mehryar Mohri, Yutao Zhong

Abstract

Training large-scale generative models is resource-intensive and relies heavily on heuristic dataset weighting. We address two fundamental questions: Can we train Large Language Models (LLMs) modularly-combining small, domain-specific experts to match monolithic performance-and can we do so robustly for any data mixture, eliminating heuristic tuning? We present a theoretical framework for modular generative modeling where a set of pre-trained experts are combined via a gating mechanism. We define the space of normalized gating functions, $G_{1}$, and formulate the problem as a minimax game to find a single robust gate that minimizes divergence to the worst-case data mixture. We prove the existence of such a robust gate using Kakutani's fixed-point theorem and show that modularity acts as a strong regularizer, with generalization bounds scaling with the lightweight gate's complexity. Furthermore, we prove that this modular approach can theoretically outperform models retrained on aggregate data, with the gap characterized by the Jensen-Shannon Divergence. Finally, we introduce a scalable Stochastic Primal-Dual algorithm and a Structural Distillation method for efficient inference. Empirical results on synthetic and real-world datasets confirm that our modular architecture effectively mitigates gradient conflict and can robustly outperform monolithic baselines.

Metadata

arXiv ID: 2602.17554
Provider: ARXIV
Primary Category: cs.LG
Published: 2026-02-19
Fetched: 2026-02-21 18:51

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2602.17554v1</id>\n    <title>A Theoretical Framework for Modular Learning of Robust Generative Models</title>\n    <updated>2026-02-19T17:09:13Z</updated>\n    <link href='https://arxiv.org/abs/2602.17554v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2602.17554v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Training large-scale generative models is resource-intensive and relies heavily on heuristic dataset weighting. We address two fundamental questions: Can we train Large Language Models (LLMs) modularly-combining small, domain-specific experts to match monolithic performance-and can we do so robustly for any data mixture, eliminating heuristic tuning? We present a theoretical framework for modular generative modeling where a set of pre-trained experts are combined via a gating mechanism. We define the space of normalized gating functions, $G_{1}$, and formulate the problem as a minimax game to find a single robust gate that minimizes divergence to the worst-case data mixture. We prove the existence of such a robust gate using Kakutani's fixed-point theorem and show that modularity acts as a strong regularizer, with generalization bounds scaling with the lightweight gate's complexity. Furthermore, we prove that this modular approach can theoretically outperform models retrained on aggregate data, with the gap characterized by the Jensen-Shannon Divergence. Finally, we introduce a scalable Stochastic Primal-Dual algorithm and a Structural Distillation method for efficient inference. Empirical results on synthetic and real-world datasets confirm that our modular architecture effectively mitigates gradient conflict and can robustly outperform monolithic baselines.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.LG'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='stat.ML'/>\n    <published>2026-02-19T17:09:13Z</published>\n    <arxiv:primary_category term='cs.LG'/>\n    <author>\n      <name>Corinna Cortes</name>\n    </author>\n    <author>\n      <name>Mehryar Mohri</name>\n    </author>\n    <author>\n      <name>Yutao Zhong</name>\n    </author>\n  </entry>"
}