一个模型通过重分配logits实现行为调控，无需重新训练。
AI Model Modulation with Logits Redistribution
- 通过logits重分配实现无需重训练的模型行为调节。
- 支持输出质量动态控制与输入特征聚焦切换，跨任务通用。
- 适用于图像分类、分割、文本生成等多场景，兼容主流架构。
大规模模型通常需适应不同所有者和用户的需求，但维护多个专用版本效率低下。为此，我们提出AIM，一种新型模型调制范式，使单一模型能根据具体终端需求表现出多样化行为。AIM支持两种关键调制模式：效用调制和聚焦调制。前者让模型所有者动态控制输出质量以提供不同效用水平，后者让用户精确控制模型关注的输入特征。AIM引入了一种训练数据无关且无需重训练的logits重分配策略。我们基于联合概率分布对logits排序的统计特性，建立了AIM调控能力的理论基础。评估结果证实AIM在实际应用中的可行性与通用性，涵盖图像分类、语义分割和文本生成等任务，以及ResNet、SegFormer和Llama等主流架构。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。