AI Model Modulation with Logits Redistribution

2026-03-13

AI Summary

This work addresses the challenge of adapting large AI models to diverse application requirements without incurring the high cost of maintaining multiple specialized versions. To this end, the authors propose AIM (Adaptive Inference Modulation), a novel paradigm that enables unified multi-behavior modulation during inference—without requiring any training data or model retraining. AIM leverages a logits redistribution strategy to support both utility modulation (dynamically adjusting output quality) and focus modulation (steering attention toward specific input features), with theoretical guarantees grounded in the joint probability distribution derived from logits ranking. Extensive experiments demonstrate that AIM is effective across image classification, semantic segmentation, and text generation tasks, and is compatible with mainstream architectures including ResNet, SegFormer, and Llama.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Large Vision Models
• Humans and AI: Human-Aware Planning and Behavior Prediction

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: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Large-scale models are typically adapted to meet the diverse requirements of model owners and users. However, maintaining multiple specialized versions of the model is inefficient. In response, we propose AIM, a novel model modulation paradigm that enables a single model to exhibit diverse behaviors to meet the specific end requirements. AIM enables two key modulation modes: utility and focus modulations. The former provides model owners with dynamic control over output quality to deliver varying utility levels, and the latter offers users precise control to shift model's focused input features. AIM introduces a logits redistribution strategy that operates in a training data-agnostic and retraining-free manner. We establish a formal foundation to ensure AIM's regulation capability, based on the statistical properties of logits ordering via joint probability distributions. Our evaluation confirms AIM's practicality and versatility for Al model modulation, with tasks spanning image classification, semantic segmentation and text generation, and prevalent architectures including ResNet, SegFormer and Llama.

Problem

• model modulation
• large-scale models
• logits redistribution
• utility control
• focus control

Innovation

• logits redistribution
• model modulation
• training-free adaptation
• utility modulation
• focus modulation