The paper introduces a semantic-first temporal scheduling strategy for flow-matching models in affect-controlled image generation. By using deterministic ODE sampling during the high-noise phase and restricting stochastic SDE exploration and group relative policy optimization (GRPO) to a late window, the method improves prompt faithfulness, affect accuracy, and human-rated semantic preservation. Experiments on FLUX.1-dev and SD3.5-Medium demonstrate reduced subject drift, faster GRPO training, and a clear trade-off between affect-semantic quality and computational cost, establishing denoising time as a practical control dimension for reliable emotional image generation.
arXiv:2610.33460v1 Announce Type: new Abstract: Controllable image synthesis should modify visual affect without replacing the objects, layout, or relations specified by a prompt. Existing methods ad...
arXiv:2609.12830v1 Announce Type: new Abstract: Continuous emotion control in text-to-image generation requires a model to improve affective alignment without changing the objects, layout, or scene d...
Group Relative Policy Optimization (GRPO) has advanced generative model post-training in verifiable domains such as code and mathematics, but evaluating video quality, especially emotional expression,...
EmoReward is a self‑evolving reward calibration framework for emotional video generation using reinforcement learning. It distills reliability from reward traces during training to adaptively weight multiple reward primitives—such as visual quality, motion, and text alignment—based on emotional evidence. The method employs grounding cues, heterogeneous signal primitives, reflective reward gating, and a RewardMemory module to calibrate fusion weights across training runs, demonstrating improvements on HunyuanVideo 1.5 and transferability to SD 3.5 and FLUX.
CAST introduces a reinforcement‑learning fine‑tuning framework for diffusion models that addresses three key limitations: it automatically selects the denoising window based on each model’s trajectory, decomposes prompts into verifiable semantic atoms via Causal Scene Graphs, and applies atom‑level rewards spatially weighted in the policy objective. The method is applied to FLUX.2‑dev and Qwen‑Image‑2512, yielding up to 3.07× improvement on the hardest GenEval 2 prompts compared with Flow‑GRPO while also enhancing overall generation quality.
PreviewDiff is a test‑time search method that uses multimodal critics to guide diffusion model sampling. By decoding partial previews at selected denoising checkpoints, scoring them with a multimodal judge, and branching over semantic prompt edits, it allows the generation process to be edited and rerouted before completion. The approach consistently outperforms budget‑matched Best‑of‑N sampling and scalar‑search baselines on image and video benchmarks, with early interventions and wider search yielding the biggest gains.
arXiv:2609.38329v1 Announce Type: new Abstract: Group-relative RL methods such as Flow-GRPO post-train image generators by exploring with isotropic Gaussian noise added at every denoising step. This...
HorizonComposer is a new instruction‑guided video editing framework that preserves traffic semantics in driving videos. It uses a hybrid dataset of real and synthetic edits, applies advantage conditioning with dense pixel‑level rewards during training, and introduces thinking frames to reduce temporal shock. Experiments show state‑of‑the‑art spatiotemporal consistency, improved content preservation, and gains in downstream road line segmentation and multi‑object detection.
arXiv:2606.04457v2 Announce Type: replace Abstract: Incorporating visual semantic representations as an intermediate step before image generation can reduce the modeling difficulty between text and i...
arXiv:2607. 10165v1 Announce Type: cross Abstract: Emotion-aware artistic image generation requires an image to match the input prompt, follow the specified artistic style, and convey the target emotion.
AffectDelta is a new image editing framework that moves beyond single emotion labels by modeling edits as transitions between eight‑dimensional emotion distributions. It uses a frozen Emotion Distribution Predictor to estimate the source image’s affective state and encodes the signed difference to guide a diffusion backbone that applies context‑dependent semantic and appearance changes. The authors created a large AffectPair‑249K dataset of source‑target pairs and show that AffectDelta outperforms six baselines in both affective alignment and content preservation, with ablation studies supporting their design choices.
SCOPE introduces a reinforcement learning approach for text‑to‑image prompt engineering that respects semantic spans such as objects, counts, attributes, and relations. By pooling policy log‑probability shifts within each semantic span, the method assigns a shared coefficient to token losses while using a single advantage per completion, eliminating the need for separate span rewards or extra image renders. Experiments on Qwen‑Image, SD3, and FLUX.1‑dev show average relative gains of 9.48% on GeneVal and 7.62% on T2I‑CompBench compared to vanilla GRPO, with ablation studies confirming the benefit of semantic boundaries and pooling.