AI Model Modulation with Logits Redistribution
This addresses the problem for model owners and users by providing dynamic control over model behavior, though it is incremental as it builds on existing modulation concepts.
The paper tackles the inefficiency of maintaining multiple specialized versions of large-scale AI models by proposing AIM, a novel model modulation paradigm that enables a single model to exhibit diverse behaviors for varying utility levels and focused input features, achieving practical results across tasks like image classification, semantic segmentation, and text generation without retraining.
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.