3D Gaussian splatting (3DGS) is a popular radiance field method, with many application-specific extensions. Most variants rely on the same core algorithm: depth-sorting of Gaussian splats then rasterizing in primitive order. This ensures correct alpha compositing, but can cause rendering artifacts due to built-in approximations. Moreover, for a fixed representation, sorted rendering offers little control over render cost and visual fidelity. For example, and counter-intuitively, rendering a lower-resolution image is not necessarily faster. In this work, we address the above limitations by combining 3D Gaussian splatting with stochastic rasterization. Concretely, we leverage an unbiased Monte Carlo estimator of the volume rendering equation. This removes the need for sorting, and allows for accurate 3D blending of overlapping Gaussians. The number of Monte Carlo samples further imbues 3DGS with a way to trade off computation time and quality. We implement our method using OpenGL shaders, enabling efficient rendering on modern GPU hardware. At a reasonable visual quality, our method renders more than four times faster than sorted rasterization.

Text-to-image generative models often struggle with long prompts detailing complex scenes, diverse objects with distinct visual characteristics and spatial relationships. In this work, we propose SCoPE (Scheduled interpolation of Coarse-to-fine Prompt Embeddings), a training-free method to improve text-to-image alignment by progressively refining the input prompt in a coarse-to-fine-grained manner. Given a detailed input prompt, we first decompose it into multiple sub-prompts which evolve from describing broad scene layout to highly intricate details. During inference, we interpolate between these sub-prompts and thus progressively introduce finer-grained details into the generated image. Our training-free plug-and-play approach significantly enhances prompt alignment, achieves an average improvement of up to +4% in Visual Question Answering (VQA) scores over the Stable Diffusion baselines on 85% of the prompts from the GenAI-Bench dataset.

Domain Generalization aims to develop models that can generalize to novel and unseen data distributions. In this work, we study how model architectures and pre-training objectives impact feature richness and propose a method to effectively leverage them for domain generalization. Specifically, given a pre-trained feature space, we first discover latent domain structures, referred to as pseudo-domains, that capture domain-specific variations in an unsupervised manner. Next, we augment existing classifiers with these complementary pseudo-domain representations making them more amenable to diverse unseen test domains. We analyze how different pre-training feature spaces differ in the domain-specific variances they capture. Our empirical studies reveal that features from diffusion models excel at separating domains in the absence of explicit domain labels and capture nuanced domain-specific information. On 5 datasets, we show that our very simple framework improves generalization to unseen domains by a maximum test accuracy improvement of over 4% compared to the standard baseline Empirical Risk Minimization (ERM). Crucially, our method outperforms most algorithms that access domain labels during training.

Despite the remarkable progress in text-to-image generative models, they are prone to adversarial attacks and inadvertently generate unsafe, unethical content. Existing approaches often rely on fine-tuning models to remove specific concepts, which is computationally expensive, lack scalability, and/or compromise generation quality. In this work, we propose a novel framework leveraging k-sparse autoencoders (k-SAEs) to enable efficient and interpretable concept manipulation in diffusion models. Specifically, we first identify interpretable monosemantic concepts in the latent space of text embeddings and leverage them to precisely steer the generation away or towards a given concept (e.g., nudity) or to introduce a new concept (e.g., photographic style). Through extensive experiments, we demonstrate that our approach is very simple, requires no retraining of the base model nor LoRA adapters, does not compromise the generation quality, and is robust to adversarial prompt manipulations. Our method yields an improvement of 20.01% in unsafe concept removal, is effective in style manipulation, and is ∼5x faster than current state-of-the-art.

We study rich visual semantic information is represented within various layers and denoising timesteps of different diffusion architectures. We uncover monosemantic interpretable features by leveraging k-sparse autoencoders (k-SAE). We substantiate our mechanistic interpretations via transfer learning using light-weight classifiers on off-the-shelf diffusion models' features. On datasets, we demonstrate the effectiveness of diffusion features for representation learning. We provide in-depth analysis of how different diffusion architectures, pre-training datasets, and language model conditioning impacts visual representation granularity, inductive biases, and transfer learning capabilities. Our work is a critical step towards deepening interpretability of black-box diffusion models. Code and visualizations available at: https://github.com/revelio-diffusion/revelio

State-of-the-art generative text-to-image models are known to exhibit social biases and over-represent certain groups like people of perceived lighter skin tones and men in their outcomes. In this work, we propose a method to mitigate such biases and ensure that the outcomes are fair across different groups of people. We do this by finetuning text-to-image models on synthetic data that varies in perceived skin tones and genders constructed from diverse text prompts. These text prompts are constructed from multiplicative combinations of ethnicities, genders, professions, age groups, and so on, resulting in diverse synthetic data. Our diversity finetuned (DFT) model improves the group fairness metric by 150% for perceived skin tone and 97.7% for perceived gender. Compared to baselines, DFT models generate more people with perceived darker skin tone and more women. To foster open research, we will release all text prompts and code to generate training images.

Text-guided generative diffusion models unlock powerful image creation and editing tools. While these have been extended to video generation, current approaches that edit the content of existing footage while retaining structure require expensive re-training for every input or rely on error-prone propagation of image edits across frames. In this work, we present a structure and content-guided video diffusion model that edits videos based on visual or textual descriptions of the desired output. Conflicts between user-provided content edits and structure representations occur due to insufficient disentanglement between the two aspects. As a solution, we show that training on monocular depth estimates with varying levels of detail provides control over structure and content fidelity. Our model is trained jointly on images and videos which also exposes explicit control of temporal consistency through a novel guidance method. Our experiments demonstrate a wide variety of successes; fine-grained control over output characteristics, customization based on a few reference images, and a strong user preference towards results by our model.

Many video editing tasks such as rotoscoping or object removal require the propagation of context across frames. While transformers and other attention-based approaches that aggregate features globally have demonstrated great success at propagating object masks from keyframes to the whole video, they struggle to propagate high-frequency details such as textures faithfully. We hypothesize that this is due to an inherent bias of global attention towards low-frequency features. To overcome this limitation, we present a two-stream approach, where high-frequency features interact locally and low-frequency features interact globally. The global interaction stream remains robust in difficult situations such as large camera motions, where explicit alignment fails. The local interaction stream propagates high-frequency details through deformable feature aggregation and, informed by the global interaction stream, learns to detect and correct errors of the deformation field. We evaluate our two-stream approach for inpainting tasks, where experiments show that it improves both the propagation of features within a single frame as required for image inpainting, as well as their propagation from keyframes to target frames. Applied to video inpainting, our approach leads to 44% and 26% improvements in FID and LPIPS scores.

By decomposing the image formation process into a sequential application of denoising autoencoders, diffusion models (DMs) achieve state-of-the-art synthesis results on image data and beyond. Additionally, their formulation allows for a guiding mechanism to control the image generation process without retraining. However, since these models typically operate directly in pixel space, optimization of powerful DMs often consumes hundreds of GPU days and inference is expensive due to sequential evaluations. To enable DM training on limited computational resources while retaining their quality and flexibility, we apply them in the latent space of powerful pretrained autoencoders. In contrast to previous work, training diffusion models on such a representation allows for the first time to reach a near-optimal point between complexity reduction and detail preservation, greatly boosting visual fidelity. By introducing cross-attention layers into the model architecture, we turn diffusion models into powerful and flexible generators for general conditioning inputs such as text or bounding boxes and high-resolution synthesis becomes possible in a convolutional manner. Our latent diffusion models (LDMs) achieve a new state of the art for image inpainting and highly competitive performance on various tasks, including unconditional image generation, semantic scene synthesis, and super-resolution, while significantly reducing computational requirements compared to pixel-based DMs.

In the art of video editing, sound helps add character to an object and immerse the viewer within a space. Through formative interviews with professional editors (N=10), we found that the task of adding sounds to video can be challenging. This paper presents Soundify, a system that assists editors in matching sounds to video. Given a video, Soundify identifies matching sounds, synchronizes the sounds to the video, and dynamically adjusts panning and volume to create spatial audio. In a human evaluation study (N=889), we show that Soundify is capable of matching sounds to video out-of-the-box for a diverse range of audio categories. In a within-subjects expert study (N=12), we demonstrate the usefulness of Soundify in helping video editors match sounds to video with lighter workload, reduced task completion time, and improved usability.