Palette of Language Models: A Solver for Controlled Text Generation

2025-03-14

AI Summary

In multi-attribute controllable text generation, prompt design is challenging, and attribute overlap/conflict renders linear combination ineffective. Method: This paper proposes a novel language model composition paradigm grounded in the law of total probability and conditional mutual information minimization. We introduce the first “palette”-inspired compositional theory, analogizing attribute intensities to color channels and formalizing their mapping to generative style; we rigorously prove two theoretical properties—positive correlation and attribute enhancement—to enable interpretable, principled composition. Our approach integrates probabilistic graphical modeling, conditional mutual information optimization, single/multi-attribute fine-tuning, and collaborative prompt-space modeling. Results: Experiments demonstrate significant improvements over linear baselines across both single- and multi-attribute control tasks, achieving consistent gains in attribute consistency, text quality, and control precision.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Computational Creativity

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Recent advancements in large language models have revolutionized text generation with their remarkable capabilities. These models can produce controlled texts that closely adhere to specific requirements when prompted appropriately. However, designing an optimal prompt to control multiple attributes simultaneously can be challenging. A common approach is to linearly combine single-attribute models, but this strategy often overlooks attribute overlaps and can lead to conflicts. Therefore, we propose a novel combination strategy inspired by the Law of Total Probability and Conditional Mutual Information Minimization on generative language models. This method has been adapted for single-attribute control scenario and is termed the Palette of Language Models due to its theoretical linkage between attribute strength and generation style, akin to blending colors on an artist's palette. Moreover, positive correlation and attribute enhancement are advanced as theoretical properties to guide a rational combination strategy design. We conduct experiments on both single control and multiple control settings, and achieve surpassing results.

Problem

• Challenges in controlling multiple text attributes simultaneously.
• Linear combination of single-attribute models causes conflicts.
• Proposes a novel strategy for controlled text generation.

Innovation

• Combines models using Law of Total Probability
• Minimizes Conditional Mutual Information for control
• Links attribute strength to generation style