Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ProCal: Probability Calibration for Neighborhood-Guided Source-Free Domain Adaptation

Ying Zheng, Yiyi Zhang, Yi Wang, Lap-Pui Chau

Abstract

Source-Free Domain Adaptation (SFDA) adapts pre-trained models to unlabeled target domains without requiring access to source data. Although state-of-the-art methods leveraging local neighborhood structures show promise for SFDA, they tend to over-rely on prediction similarity among neighbors. This over-reliance accelerates the forgetting of source knowledge and increases susceptibility to local noise overfitting. To address these issues, we introduce ProCal, a probability calibration method that dynamically calibrates neighborhood-based predictions through a dual-model collaborative prediction mechanism. ProCal integrates the source model's initial predictions with the current model's online outputs to effectively calibrate neighbor probabilities. This strategy not only mitigates the interference of local noise but also preserves the discriminative information from the source model, thereby achieving a balance between knowledge retention and domain adaptation. Furthermore, we design a joint optimization objective that combines a soft supervision loss with a diversity loss to guide the target model. Our theoretical analysis shows that ProCal converges to an equilibrium where source knowledge and target information are effectively fused, reducing both knowledge forgetting and overfitting. We validate the effectiveness of our approach through extensive experiments on 31 cross-domain tasks across four public datasets. Our code is available at: https://github.com/zhengyinghit/ProCal.

ProCal: Probability Calibration for Neighborhood-Guided Source-Free Domain Adaptation

Abstract

Source-Free Domain Adaptation (SFDA) adapts pre-trained models to unlabeled target domains without requiring access to source data. Although state-of-the-art methods leveraging local neighborhood structures show promise for SFDA, they tend to over-rely on prediction similarity among neighbors. This over-reliance accelerates the forgetting of source knowledge and increases susceptibility to local noise overfitting. To address these issues, we introduce ProCal, a probability calibration method that dynamically calibrates neighborhood-based predictions through a dual-model collaborative prediction mechanism. ProCal integrates the source model's initial predictions with the current model's online outputs to effectively calibrate neighbor probabilities. This strategy not only mitigates the interference of local noise but also preserves the discriminative information from the source model, thereby achieving a balance between knowledge retention and domain adaptation. Furthermore, we design a joint optimization objective that combines a soft supervision loss with a diversity loss to guide the target model. Our theoretical analysis shows that ProCal converges to an equilibrium where source knowledge and target information are effectively fused, reducing both knowledge forgetting and overfitting. We validate the effectiveness of our approach through extensive experiments on 31 cross-domain tasks across four public datasets. Our code is available at: https://github.com/zhengyinghit/ProCal.
Paper Structure (13 sections, 1 theorem, 19 equations, 4 figures, 12 tables, 1 algorithm)

This paper contains 13 sections, 1 theorem, 19 equations, 4 figures, 12 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Method
  4. Problem Definition
  5. Critical Analysis of Existing Methods
  6. The ProCal Mechanism
  7. Theoretical Analysis
  8. Experiments
  9. Implementation Details
  10. Comparative Results
  11. Ablation and Backbone Analysis
  12. Further Analysis and Discussion
  13. Conclusion

Key Result

Theorem 1

Under the probability simplex constraint, the fixed-point solution for the sample $\bm{x}_i$ is given by $\bm{p}_i^* = \frac{1}{2\gamma}\left( \frac{2\gamma + \bm{1}^\top \bm{q}_i}{C}\,\bm{1} - \bm{q}_i \right)$, where $\bm{1}$ denotes the all-ones vector in $\mathbb{R}^C$.

Figures (4)

  • Figure 1: Illustration of the effect of neighborhood distribution calibration. (a) Direct optimization with neighborhood constraints may lead to source knowledge forgetting and overfitting to noisy local target structures. (b) Neighborhood distribution calibration mitigates these issues and yields better generalization.
  • Figure 2: Training dynamics on the Ar$\rightarrow$Cl task of Office-Home. Compared with SHOT and AaD, ProCal consistently exhibits a lower source forgetting rate and a lower incorrect supervision rate throughout training, while achieving the highest target accuracy. This suggests that ProCal better preserves transferable source knowledge and reduces harmful target supervision during adaptation.
  • Figure 3: Feature space visualizations of the representations learned by (a) the source model, (b) AaD yang2022attracting, and (c) our ProCal on the Ar$\rightarrow$Cl task of Office-Home. ProCal produces more compact and better-separated class clusters.
  • Figure 4: Sensitivity analysis of the hyperparameters $\gamma_1$, $\beta_1$, and $k$ on Office-31, VisDA-C, and Office-Home. Although the performance varies to some extent with different hyperparameter choices, it remains relatively stable in the vicinity of the selected values, suggesting that the proposed method is not overly sensitive to hyperparameter tuning.

Theorems & Definitions (1)

  • Theorem 1