SliderSpace: Decomposing the Visual Capabilities of Diffusion Models
This work addresses the need for more intuitive and efficient control over diffusion models for users in creative and AI research domains, though it is incremental as it builds on existing control methods.
The authors tackled the problem of decomposing the visual capabilities of diffusion models into controllable directions, resulting in a framework that automatically discovers multiple interpretable directions from a single text prompt, with user studies validating more diverse and useful variations compared to baselines.
We present SliderSpace, a framework for automatically decomposing the visual capabilities of diffusion models into controllable and human-understandable directions. Unlike existing control methods that require a user to specify attributes for each edit direction individually, SliderSpace discovers multiple interpretable and diverse directions simultaneously from a single text prompt. Each direction is trained as a low-rank adaptor, enabling compositional control and the discovery of surprising possibilities in the model's latent space. Through extensive experiments on state-of-the-art diffusion models, we demonstrate SliderSpace's effectiveness across three applications: concept decomposition, artistic style exploration, and diversity enhancement. Our quantitative evaluation shows that SliderSpace-discovered directions decompose the visual structure of model's knowledge effectively, offering insights into the latent capabilities encoded within diffusion models. User studies further validate that our method produces more diverse and useful variations compared to baselines. Our code, data and trained weights are available at https://sliderspace.baulab.info