Self-control: A Better Conditional Mechanism for Masked Autoregressive Model

TL;DR

This paper introduces a novel continuous masked autoregressive model with a self-control network that improves image quality and conditional control by unifying multimodal information without relying on vector quantization.

Contribution

It proposes a self-control network for continuous masked autoregressive models that enhances image quality and multimodal conditional control without vector quantization.

Findings

Improved image quality in autoregressive models.

Enhanced conditional control with multimodal data.

Unified space for conditional and generative information.

Abstract

Autoregressive conditional image generation algorithms are capable of generating photorealistic images that are consistent with given textual or image conditions, and have great potential for a wide range of applications. Nevertheless, the majority of popular autoregressive image generation methods rely heavily on vector quantization, and the inherent discrete characteristic of codebook presents a considerable challenge to achieving high-quality image generation. To address this limitation, this paper introduces a novel conditional introduction network for continuous masked autoregressive models. The proposed self-control network serves to mitigate the negative impact of vector quantization on the quality of the generated images, while simultaneously enhancing the conditional control during the generation process. In particular, the self-control network is constructed upon a continuous mask autoregressive generative model, which incorporates multimodal conditional information, including text and images, into a unified autoregressive sequence in a serial manner. Through a self-attention mechanism, the network is capable of generating images that are controllable based on specific conditions. The self-control network discards the conventional cross-attention-based conditional fusion mechanism and effectively unifies the conditional and generative information within the same space, thereby facilitating more seamless learning and fusion of multimodal features.