# Von Mises-Fisher Mixture Model with Dynamic Shrinkage for Realistic Test-Time Transduction

arXiv: https://arxiv.org/abs/2607.15851 · HTML: https://mlsanity.com/paper/2607.15851/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Prompt learning and adaptation of CLIP, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 14 of its 60 references that are in the map.
- First posted on arXiv 17 Jul 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Prompt learning and adaptation of CLIP](https://mlsanity.com/mllm/prompt-learning-and-adaptation-of-clip/)

## Builds on (most cited first, 14 of 14)

- [Fine-Grained Visual Classification of Aircraft](https://mlsanity.com/paper/1306.5151/) (2013)
- [EVA-CLIP: Improved Training Techniques for CLIP at Scale](https://mlsanity.com/paper/2303.15389/) (2023)
- [The Linear Representation Hypothesis and the Geometry of Large Language Models](https://mlsanity.com/paper/2311.03658/) (2023)
- [Tent: Fully Test-time Adaptation by Entropy Minimization](https://mlsanity.com/paper/2006.10726/) (2020)
- [On Pitfalls of Test-Time Adaptation](https://mlsanity.com/paper/2306.03536/) (2023)
- [A Hard-to-Beat Baseline for Training-free CLIP-based Adaptation](https://mlsanity.com/paper/2402.04087/) (2024)
- [von Mises-Fisher Mixture Model-based Deep learning: Application to Face Verification](https://mlsanity.com/paper/1706.04264/) (2017)
- [DOTA: Distributional Test-Time Adaptation of Vision-Language Models](https://mlsanity.com/paper/2409.19375/) (2024)
- [BoostAdapter: Improving Vision-Language Test-Time Adaptation via Regional Bootstrapping](https://mlsanity.com/paper/2410.15430/) (2024)
- [DINO as a von Mises-Fisher mixture model](https://mlsanity.com/paper/2405.10939/) (2024)
- [Backpropagation-Free Test-Time Adaptation via Probabilistic Gaussian Alignment](https://mlsanity.com/paper/2508.15568/) (2025)
- [What Drives Test-Time Adaptation for CLIP? A Controlled Empirical Study from an Update Perspective](https://mlsanity.com/paper/2606.14299/) (2026)
- [Do All Individual Layers Help? An Empirical Study of Task-Interfering Layers in Vision-Language Models](https://mlsanity.com/paper/2602.01167/) (2026)
- [Adaptive Disentangled Representation Learning for Incomplete Multi-View Multi-Label Classification](https://mlsanity.com/paper/2601.05785/) (2026)

## Abstract (arXiv)

A range of methods aim to enhance the performance of vision-language models (VLMs) at test time. Among them, transduction has emerged as a promising paradigm due to its strong compatibility and efficiency. However, realistic evaluations often involve highly imbalanced class distributions, which cause performance degradation or even collapse. In this work, we systematically revisit transduction from the perspective of penalized likelihood estimation (PLE), showing that PLE with a KL-divergence anchor term naturally yields an adaptive shrinkage behavior between prior anchors and empirical estimates. From this viewpoint, the brittleness of transductive methods can be attributed to the absence of anchoring mechanism and static modeling of the shrinkage strength. Therefore, we propose Mixture of Von Mises-Fisher Models with Dynamic Shrinkage (MOON). MOON is built upon a mixture of von Mises-Fisher distributions to model feature representations on the unit hypersphere. To handle imbalance, MOON dynamically adjusts the shrinkage strength using zero-shot priors at both instance and class levels. Thus, it suppresses unreliable assignments and prevents harmful updates from outlier classes, thereby mitigating negative transfer. MOON is model-agnostic, training-free, and requires no task-specific hyperparameter tuning. Extensive experiments further validate the advantage of MOON in both performance and efficiency. Our code is available at https://github.com/walawalagoose/MOON