# DT-UFC: Universal Large Model Feature Coding via Peaky-to-Balanced Distribution Transformation

arXiv: https://arxiv.org/abs/2506.16495 · HTML: https://mlsanity.com/paper/2506.16495/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 28 of its 65 references that are in the map.
- First posted on arXiv 19 Jun 2025.

## Builds on (most cited first, 28 of 28)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Scaling Rectified Flow Transformers for High-Resolution Image Synthesis](https://mlsanity.com/paper/2403.03206/) (2024)
- [Qwen Technical Report](https://mlsanity.com/paper/2309.16609/) (2023)
- [Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context](https://mlsanity.com/paper/2403.05530/) (2024)
- [Learning Transferable Features with Deep Adaptation Networks](https://mlsanity.com/paper/1502.02791/) (2015)
- [Deep CORAL: Correlation Alignment for Deep Domain Adaptation](https://mlsanity.com/paper/1607.01719/) (2016)
- [A Closer Look at Memorization in Deep Networks](https://mlsanity.com/paper/1706.05394/) (2017)
- [Variational image compression with a scale hyperprior](https://mlsanity.com/paper/1802.01436/) (2018)
- [OpenFedLLM: Training Large Language Models on Decentralized Private Data via Federated Learning](https://mlsanity.com/paper/2402.06954/) (2024)
- [USTC-TD: A Test Dataset and Benchmark for Image and Video Coding in 2020s](https://mlsanity.com/paper/2409.08481/) (2024)
- [TransTIC: Transferring Transformer-based Image Compression from Human Perception to Machine Perception](https://mlsanity.com/paper/2306.05085/) (2023)
- [Image Coding for Machines with Omnipotent Feature Learning](https://mlsanity.com/paper/2207.01932/) (2022)
- [Towards Analysis-friendly Face Representation with Scalable Feature and Texture Compression](https://mlsanity.com/paper/2004.10043/) (2020)
- [Structured Multi-modal Feature Embedding and Alignment for Image-Sentence Retrieval](https://mlsanity.com/paper/2108.02417/) (2021)
- [Object Segmentation-Assisted Inter Prediction for Versatile Video Coding](https://mlsanity.com/paper/2403.11694/) (2024)
- [Multi-task learning with compressible features for Collaborative Intelligence](https://mlsanity.com/paper/1902.05179/) (2019)
- [A Coding Framework and Benchmark towards Low-Bitrate Video Understanding](https://mlsanity.com/paper/2202.02813/) (2022)
- [Feature Alignment by Uncertainty and Self-Training for Source-Free Unsupervised Domain Adaptation](https://mlsanity.com/paper/2208.14888/) (2022)
- [Bit Allocation for Multi-Task Collaborative Intelligence](https://mlsanity.com/paper/2002.07048/) (2020)
- [End-to-End Learnable Multi-Scale Feature Compression for VCM](https://mlsanity.com/paper/2306.16670/) (2023)
- [Feature Coding in the Era of Large Models: Dataset, Test Conditions, and Benchmark](https://mlsanity.com/paper/2412.04307/) (2024)
- [Rate-Distortion-Cognition Controllable Versatile Neural Image Compression](https://mlsanity.com/paper/2407.11700/) (2024)
- [Free-VSC: Free Semantics from Visual Foundation Models for Unsupervised Video Semantic Compression](https://mlsanity.com/paper/2409.11718/) (2024)
- [DMOFC: Discrimination Metric-Optimized Feature Compression](https://mlsanity.com/paper/2405.04044/) (2024)
- [Toward Scalable Image Feature Compression: A Content-Adaptive and Diffusion-Based Approach](https://mlsanity.com/paper/2410.06149/) (2024)
- [Unified Coding for Both Human Perception and Generalized Machine Analytics with CLIP Supervision](https://mlsanity.com/paper/2501.04579/) (2025)

## Built on by (most cited first, 6 of 6)

- [Just Noticeable Difference Modeling for Deep Visual Features](https://mlsanity.com/paper/2601.21933/) (2026)
- [Machines Serve Human: A Novel Variable Human-machine Collaborative Compression Framework](https://mlsanity.com/paper/2511.08915/) (2025)
- [Towards Large Model Feature Coding](https://mlsanity.com/paper/2605.24025/) (2026)
- [Visual Token Codec: Unleashing Spatial Redundancy for ViT Feature Coding](https://mlsanity.com/paper/2608.08832/) (2026)
- [Just Noticeable Difference Modeling for Token Compression in Vision-Language-Action Models](https://mlsanity.com/paper/2608.21247/) (2026)
- [On Task Scope and Information Retention in Source Coding](https://mlsanity.com/paper/2609.37575/) (2026)

## Abstract (arXiv)

Like image coding in visual data transmission, feature coding is essential for the distributed deployment of large models by significantly reducing transmission and storage burden. However, prior studies have mostly targeted task- or model-specific scenarios, leaving the challenge of universal feature coding across diverse large models largely unexplored. In this paper, we present the first systematic study on universal feature coding for large models. The key challenge lies in the inherently diverse and distributionally incompatible nature of features extracted from different models. For example, features from DINOv2 exhibit highly peaky, concentrated distributions, while those from Stable Diffusion 3 (SD3) are more dispersed and uniform. This distributional heterogeneity severely hampers both compression efficiency and cross-model generalization. To address this, we propose a learned peaky-to-balanced distribution transformation, which reshapes highly skewed feature distributions into a common, balanced target space. This transformation is non-uniform, data-driven, and plug-and-play, enabling effective alignment of heterogeneous distributions without modifying downstream codecs. With this alignment, a universal codec trained on the balanced target distribution can effectively generalize to features from different models and tasks. We validate our approach on three representative large models (LLaMA3, DINOv2, and SD3) across multiple tasks and modalities. Extensive experiments show that our method achieves notable improvements in both compression efficiency and cross-model generalization over task-specific baselines. All source code has been made available at https://github.com/chansongoal/DT-UFC.