Analog Bits: Generating Discrete Data using Diffusion Models with Self-Conditioning
We present Bit Diffusion: a simple and generic approach for generating discrete data with continuous state and continuous time diffusion models. The main idea behind our approach is to first represent the discrete data as binary bits, and then train a continuous diffusion model to model these bits as real numbers which we call analog bits. To generate samples, the model first generates the analog bits, which are then thresholded to obtain the bits that represent the discrete variables. We further propose two simple techniques, namely Self-Conditioning and Asymmetric Time Intervals, which lead to a significant improvement in sample quality. Despite its simplicity, the proposed approach can achieve strong performance in both discrete image generation and image captioning tasks. For discrete image generation, we significantly improve previous state-of-the-art on both CIFAR-10 (which has 3K discrete 8-bit tokens) and ImageNet-64x64 (which has 12K discrete 8-bit tokens), outperforming the best autoregressive model in both sample quality (measured by FID) and efficiency. For image captioning on MS-COCO dataset, our approach achieves competitive results compared to autoregressive models.
Code (7)
Tasks
Image CaptioningImage GenerationMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Fine-Tuning Discrete Diffusion Models via Reward Optimization with Applications to DNA and Protein Design
Recent studies have demonstrated the strong empirical performance of diffusion models on discrete sequences across domains from natural language to biological sequence generation. For example, in the protein inverse fold…
Protein DesignReinforcement Learning (RL)Discrete Diffusion Probabilistic Models for Symbolic Music Generation
Denoising Diffusion Probabilistic Models (DDPMs) have made great strides in generating high-quality samples in both discrete and continuous domains. However, Discrete DDPMs (D3PMs) have yet to be applied to the domain of…
DenoisingMusic GenerationResBit: Residual Bit Vector for Categorical Values
One-hot vectors, a common method for representing discrete/categorical data, in machine learning are widely used because of their simplicity and intuitiveness. However, one-hot vectors suffer from a linear increase in di…
Image ClassificationTabular Data GenerationSplit Gibbs Discrete Diffusion Posterior Sampling
We study the problem of posterior sampling in discrete-state spaces using discrete diffusion models. While posterior sampling methods for continuous diffusion models have achieved remarkable progress, analogous methods f…
Foundations of Diffusion Models in General State Spaces: A Self-Contained Introduction
Although diffusion models now occupy a central place in generative modeling, introductory treatments commonly assume Euclidean data and seldom clarify their connection to discrete-state analogues. This article is a self-…
Gaussian Processes