# Unsupervised Discovery of Interpretable Directions in h-space of Pre-trained Diffusion Models

arXiv: https://arxiv.org/abs/2310.09912 · HTML: https://mlsanity.com/paper/2310.09912/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models.
- Builds on 35 of its 42 references that are in the map.
- First posted on arXiv 15 Oct 2023.

## Placed in

- Diffusion models / [Diffusion foundations and sampling](https://mlsanity.com/diffusion/diffusion-foundations-and-sampling/)

## Builds on (most cited first, 30 of 35)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [Denoising Diffusion Implicit Models](https://mlsanity.com/paper/2010.02502/) (2020)
- [Diffusion Models Beat GANs on Image Synthesis](https://mlsanity.com/paper/2105.05233/) (2021)
- [Deep Unsupervised Learning using Nonequilibrium Thermodynamics](https://mlsanity.com/paper/1503.03585/) (2015)
- [Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding](https://mlsanity.com/paper/2205.11487/) (2022)
- [A Style-Based Generator Architecture for Generative Adversarial Networks](https://mlsanity.com/paper/1812.04948/) (2018)
- [Adding Conditional Control to Text-to-Image Diffusion Models](https://mlsanity.com/paper/2302.05543/) (2023)
- [Auto-Encoding Variational Bayes](https://mlsanity.com/paper/1312.6114/) (2013)
- [Progressive Growing of GANs for Improved Quality, Stability, and Variation](https://mlsanity.com/paper/1710.10196/) (2017)
- [StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation](https://mlsanity.com/paper/1711.09020/) (2017)
- [Training Deep Nets with Sublinear Memory Cost](https://mlsanity.com/paper/1604.06174/) (2016)
- [DiffusionCLIP: Text-Guided Diffusion Models for Robust Image Manipulation](https://mlsanity.com/paper/2110.02711/) (2021)
- [GANSpace: Discovering Interpretable GAN Controls](https://mlsanity.com/paper/2004.02546/) (2020)
- [Diffusion Autoencoders: Toward a Meaningful and Decodable Representation](https://mlsanity.com/paper/2111.15640/) (2021)
- [Aligning Text-to-Image Models using Human Feedback](https://mlsanity.com/paper/2302.12192/) (2023)
- [Closed-Form Factorization of Latent Semantics in GANs](https://mlsanity.com/paper/2007.06600/) (2020)
- [Diffusion Models already have a Semantic Latent Space](https://mlsanity.com/paper/2210.10960/) (2022)
- [Perception Prioritized Training of Diffusion Models](https://mlsanity.com/paper/2204.00227/) (2022)
- [Unsupervised Discovery of Interpretable Directions in the GAN Latent Space](https://mlsanity.com/paper/2002.03754/) (2020)
- [Latent Space Oddity: on the Curvature of Deep Generative Models](https://mlsanity.com/paper/1710.11379/) (2017)
- [Refining Generative Process with Discriminator Guidance in Score-based Diffusion Models](https://mlsanity.com/paper/2211.17091/) (2022)
- [The Riemannian Geometry of Deep Generative Models](https://mlsanity.com/paper/1711.08014/) (2017)
- [Memory-Efficient Backpropagation Through Time](https://mlsanity.com/paper/1606.03401/) (2016)
- [Generating High Fidelity Data from Low-density Regions using Diffusion Models](https://mlsanity.com/paper/2203.17260/) (2022)
- [Unsupervised Representation Learning from Pre-trained Diffusion Probabilistic Models](https://mlsanity.com/paper/2212.12990/) (2022)
- [Metrics for Deep Generative Models](https://mlsanity.com/paper/1711.01204/) (2017)

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

- [Implicit Diffusion: Efficient Optimization through Stochastic Sampling](https://mlsanity.com/paper/2402.05468/) (2024)
- [DiffLoss: unleashing diffusion model as constraint for training image restoration network](https://mlsanity.com/paper/2406.19030/) (2024)
- [On Designing Diffusion Autoencoders for Efficient Generation and Representation Learning](https://mlsanity.com/paper/2506.00136/) (2025)
- [Unsupervised Panoptic Interpretation of Latent Spaces in GANs Using Space-Filling Vector Quantization](https://mlsanity.com/paper/2410.20573/) (2024)
- [RemEdit: Efficient Diffusion Editing with Riemannian Geometry](https://mlsanity.com/paper/2601.17927/) (2026)
- [RankT2I: A Submodular Framework for Discovering Interpretable and Diverse Semantics in Text-to-Image Models](https://mlsanity.com/paper/2608.14226/) (2026)

## Abstract (arXiv)

We propose the first unsupervised and learning-based method to identify interpretable directions in h-space of pre-trained diffusion models. Our method is derived from an existing technique that operates on the GAN latent space. Specifically, we employ a shift control module that works on h-space of pre-trained diffusion models to manipulate a sample into a shifted version of itself, followed by a reconstructor to reproduce both the type and the strength of the manipulation. By jointly optimizing them, the model will spontaneously discover disentangled and interpretable directions. To prevent the discovery of meaningless and destructive directions, we employ a discriminator to maintain the fidelity of shifted sample. Due to the iterative generative process of diffusion models, our training requires a substantial amount of GPU VRAM to store numerous intermediate tensors for back-propagating gradient. To address this issue, we propose a general VRAM-efficient training algorithm based on gradient checkpointing technique to back-propagate any gradient through the whole generative process, with acceptable occupancy of VRAM and sacrifice of training efficiency. Compared with existing related works on diffusion models, our method inherently identifies global and scalable directions, without necessitating any other complicated procedures. Extensive experiments on various datasets demonstrate the effectiveness of our method.