# The Prism Hypothesis: Harmonizing Semantic and Pixel Representations via Unified Autoencoding

arXiv: https://arxiv.org/abs/2512.19693 · HTML: https://mlsanity.com/paper/2512.19693/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- Builds on 49 of its 60 references that are in the map.
- First posted on arXiv 22 Dec 2025.

## Builds on (most cited first, 30 of 49)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis](https://mlsanity.com/paper/2307.01952/) (2023)
- [Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision](https://mlsanity.com/paper/2102.05918/) (2021)
- [Taming Transformers for High-Resolution Image Synthesis](https://mlsanity.com/paper/2012.09841/) (2020)
- [BEiT: BERT Pre-Training of Image Transformers](https://mlsanity.com/paper/2106.08254/) (2021)
- [Implicit Neural Representations with Periodic Activation Functions](https://mlsanity.com/paper/2006.09661/) (2020)
- [Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains](https://mlsanity.com/paper/2006.10739/) (2020)
- [Glow: Generative Flow with Invertible 1x1 Convolutions](https://mlsanity.com/paper/1807.03039/) (2018)
- [Generating Diverse High-Fidelity Images with VQ-VAE-2](https://mlsanity.com/paper/1906.00446/) (2019)
- [On the Spectral Bias of Neural Networks](https://mlsanity.com/paper/1806.08734/) (2018)
- [SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features](https://mlsanity.com/paper/2502.14786/) (2025)
- [Alias-Free Generative Adversarial Networks](https://mlsanity.com/paper/2106.12423/) (2021)
- [ImageBind: One Embedding Space To Bind Them All](https://mlsanity.com/paper/2305.05665/) (2023)
- [Cascaded Diffusion Models for High Fidelity Image Generation](https://mlsanity.com/paper/2106.15282/) (2021)
- [Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Prediction](https://mlsanity.com/paper/2404.02905/) (2024)
- [FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space](https://mlsanity.com/paper/2506.15742/) (2025)
- [SiT: Exploring Flow and Diffusion-based Generative Models with Scalable Interpolant Transformers](https://mlsanity.com/paper/2401.08740/) (2024)
- [data2vec: A General Framework for Self-supervised Learning in Speech, Vision and Language](https://mlsanity.com/paper/2202.03555/) (2022)
- [Language Model Beats Diffusion -- Tokenizer is Key to Visual Generation](https://mlsanity.com/paper/2310.05737/) (2023)
- [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://mlsanity.com/paper/2107.14795/) (2021)
- [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://mlsanity.com/paper/2202.11094/) (2022)
- [Generative Multimodal Models are In-Context Learners](https://mlsanity.com/paper/2312.13286/) (2023)
- [Unified-IO: A Unified Model for Vision, Language, and Multi-Modal Tasks](https://mlsanity.com/paper/2206.08916/) (2022)
- [Unified-IO 2: Scaling Autoregressive Multimodal Models with Vision, Language, Audio, and Action](https://mlsanity.com/paper/2312.17172/) (2023)
- [Diffusion Transformers with Representation Autoencoders](https://mlsanity.com/paper/2510.11690/) (2025)

## Built on by (most cited first, 22 of 22)

- [SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture](https://mlsanity.com/paper/2605.12500/) (2026)
- [Improved Baselines with Representation Autoencoders](https://mlsanity.com/paper/2605.18324/) (2026)
- [Tuna-2: Pixel Embeddings Beat Vision Encoders for Multimodal Understanding and Generation](https://mlsanity.com/paper/2604.24763/) (2026)
- [Making Reconstruction FID Predictive of Diffusion Generation FID](https://mlsanity.com/paper/2603.05630/) (2026)
- [What Matters for Diffusion-Friendly Latent Manifold? Prior-Aligned Autoencoders for Latent Diffusion](https://mlsanity.com/paper/2605.07915/) (2026)
- [LatentUM: Unleashing the Potential of Interleaved Cross-Modal Reasoning via a Latent-Space Unified Model](https://mlsanity.com/paper/2604.02097/) (2026)
- [Spectrum Matching: a Unified Perspective for Superior Diffusability in Latent Diffusion](https://mlsanity.com/paper/2603.14645/) (2026)
- [Diffusing in the Right Space: A Systematic Study of Latent Diffusability](https://mlsanity.com/paper/2606.03578/) (2026)
- [WavCube: Unifying Speech Representation for Understanding and Generation via Semantic-Acoustic Joint Modeling](https://mlsanity.com/paper/2605.06407/) (2026)
- [HYDRA: Representation Harmonized Tokenization for Multimodal Generation and Understanding](https://mlsanity.com/paper/2603.15228/) (2026)
- [Show the Signal, Hide the Noise: Spectral Forcing for Pixel-Space Diffusion](https://mlsanity.com/paper/2606.15236/) (2026)
- [Twins: Learn to Predict Unified Representations with Focal Loss](https://mlsanity.com/paper/2607.22531/) (2026)
- [KVAE: Family of Tokenizers for Multimodal Generative Models](https://mlsanity.com/paper/2608.05798/) (2026)
- [FuseReg: Regularizing Layer Fusion Mitigates the Reconstruction-Generation Gap in Representation Autoencoders](https://mlsanity.com/paper/2609.31620/) (2026)
- [PixelUMM: Encoder-Free Unified Image and Video Understanding and Generation](https://mlsanity.com/paper/2609.38597/) (2026)
- [A Tilted Seesaw: Revisiting Autoencoder Trade-off for Controllable Diffusion](https://mlsanity.com/paper/2601.21633/) (2026)
- [TAMISeg: Text-Aligned Multi-scale Medical Image Segmentation with Semantic Encoder Distillation](https://mlsanity.com/paper/2604.10912/) (2026)
- [SPAE: Spectrally Guided Autoencoder for Pretrained Visual Latents](https://mlsanity.com/paper/2608.01306/) (2026)
- [AffineTok: Semantic Affine Consistency for Diffusion-Friendly Visual Tokenizer](https://mlsanity.com/paper/2608.23864/) (2026)
- [Learning the Target Priors Before Image Translation: A Decoupled Training Paradigm for Cross-Modal Image Translation in Remote Sensing](https://mlsanity.com/paper/2608.28517/) (2026)
- [FLAT: Resampling Image and Text into 1D Flexible-Length Aligned Transmodal Tokens for Retrieval and Generation](https://mlsanity.com/paper/2609.16591/) (2026)
- [Image Tokenizers Should Be Trained Between Images](https://mlsanity.com/paper/2610.16759/) (2026)

## Abstract (arXiv)

Deep representations across modalities are inherently intertwined. In this paper, we systematically analyze the spectral characteristics of various semantic and pixel encoders. Interestingly, our study uncovers a highly inspiring and rarely explored correspondence between an encoder's feature spectrum and its functional role: semantic encoders primarily capture low-frequency components that encode abstract meaning, whereas pixel encoders additionally retain high-frequency information that conveys fine-grained detail. This heuristic finding offers a unifying perspective that ties encoder behavior to its underlying spectral structure. We define it as the Prism Hypothesis, where each data modality can be viewed as a projection of the natural world onto a shared feature spectrum, just like the prism. Building on this insight, we propose Unified Autoencoding (UAE), a model that harmonizes semantic structure and pixel details via an innovative frequency-band modulator, enabling their seamless coexistence. Extensive experiments demonstrate that UAE effectively unifies semantic abstraction and pixel-level fidelity within a single latent space, achieving state-of-the-art performance. Moreover, we show that UAE can be directly applied to pixel-space modeling, significantly improving both FID and IS over the vanilla JIT baseline. Our code is avaliable at: https://github.com/WeichenFan/UAE.