# OmniAlpha: Aligning Transparency-Aware Generation via Multi-Task Unified Reinforcement Learning

arXiv: https://arxiv.org/abs/2511.20211 · HTML: https://mlsanity.com/paper/2511.20211/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 Nov 2025.

## Built on by (most cited first, 3 of 3)

- [PromptHub: Enhancing Multi-Prompt Visual In-Context Learning with Locality-Aware Fusion, Concentration and Alignment](https://mlsanity.com/paper/2603.18891/) (2026)
- [UniWorld-Design: From Pixel Generation to Layer-Native Design](https://mlsanity.com/paper/2608.03971/) (2026)
- [PixelART: Image-to-Layer Decomposition without Latents or Text-to-Image Pretraining](https://mlsanity.com/paper/2609.22789/) (2026)

## Abstract (arXiv)

Transparency-aware generation requires modeling not only RGB appearance but also alpha-based opacity and cross-layer composition, which are essential for tasks such as image matting, object removal, layer decomposition, and multi-layer content creation. However, existing RGBA-related methods remain largely fragmented, with separate pipelines designed for individual tasks. While a unified model is desirable, supervised fine-tuning alone is insufficient, as localized regression objectives cannot directly optimize the compositional fidelity, alpha-boundary precision, and structural consistency required for high-quality RGBA generation. To address this, we propose OmniAlpha, a unified multi-task reinforcement learning framework for transparency-aware generation and manipulation. OmniAlpha combines an end-to-end alpha-aware VAE and a sequence-to-sequence Diffusion Transformer, with a bi-directional layer axis in positional encoding to jointly model multiple RGBA inputs and outputs within a single forward pass. Built on a multi-task SFT cold start, it further performs GRPO-style post-training with layer-aware rewards defined on decoded RGBA outputs, enabling direct optimization of cross-layer coherence and fine transparency details. Experiments across five categories of transparency-aware tasks show that OmniAlpha consistently outperforms its unified SFT baseline and achieves strong performance against specialized expert models, including a 9.07% relative reduction in RGB L1 on layer decomposition and 74%/68% improvements over conventional matting tools on SAD/Grad for automatic matting.