REVIEW 26 cited by
HPSv3: Towards Wide-Spectrum Human Preference Score
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
HPSv3: Towards Wide-Spectrum Human Preference Score
read the original abstract
Evaluating text-to-image generation models requires alignment with human perception, yet existing human-centric metrics are constrained by limited data coverage, suboptimal feature extraction, and inefficient loss functions. To address these challenges, we introduce Human Preference Score v3 (HPSv3). (1) We release HPDv3, the first wide-spectrum human preference dataset integrating 1.08M text-image pairs and 1.17M annotated pairwise comparisons from state-of-the-art generative models and low to high-quality real-world images. (2) We introduce a VLM-based preference model trained using an uncertainty-aware ranking loss for fine-grained ranking. Besides, we propose Chain-of-Human-Preference (CoHP), an iterative image refinement method that enhances quality without extra data, using HPSv3 to select the best image at each step. Extensive experiments demonstrate that HPSv3 serves as a robust metric for wide-spectrum image evaluation, and CoHP offers an efficient and human-aligned approach to improve image generation quality. The code and dataset are available at the HPSv3 Homepage.
Forward citations
Cited by 26 Pith papers
-
Cross-Space Distillation: Teaching One-Step Students with Modern Diffusion Teachers
Introduces a Bridge latent interface that maps mismatched student latents into teacher space, enabling distillation from modern diffusion teachers to compact one-step students and raising SD 1.5 HPSv3 from 5.4 to 9.4 ...
-
WBench: A Comprehensive Multi-turn Benchmark for Interactive Video World Model Evaluation
WBench is a benchmark with 289 test cases and 1,058 turns for evaluating interactive world models using 22 automated metrics validated against human judgments.
-
Towards Anatomically Plausible Human Image Generation via Synthetic Localized Preferences
ASAP generates over 10K synthetic anatomical preference pairs via targeted degradation of high-fidelity images and applies a localized margin-bounded DPO to reduce anatomical errors in text-to-image human generation, ...
-
KVPO: ODE-Native GRPO for Autoregressive Video Alignment via KV Semantic Exploration
KVPO aligns streaming autoregressive video generators with human preferences via ODE-native GRPO, using KV cache for semantic exploration and TVE for velocity-based policy modeling, yielding gains in quality and alignment.
-
Long-Text-to-Image Generation via Compositional Prompt Decomposition
PRISM lets pre-trained text-to-image models handle long prompts by breaking them into compositional parts, predicting noise separately, and merging outputs via energy-based conjunction, matching fine-tuned models whil...
-
Latent Reward Registers for Diffusion Preference Alignment
Prepending learnable register tokens to a frozen DiT yields dense latent reward estimates, enabling faster on-policy distillation and training-free guided sampling that improve preference alignment.
-
Learning Sampling Parameters for Diffusion Models
An LLM policy trained with GRPO can emit prompt-conditioned, timestep-varying diffusion sampling parameters that beat fixed defaults and prior LLM schedulers on preference metrics.
-
DetailAnywhere: Fashion Detail Generation via Cross-Modal Feature Alignment Distillation
Formalizes Fashion Detail Generation task, releases FDBench benchmark with 40K+ pairs, and proposes CFAD distillation method plus RL consistency reward that outperforms open-source baselines.
-
PortraitGen: Exemplar-Driven GRPO with Dual-Reward Guidance for Photorealistic Portrait Generation
PortraitGen integrates real-image exemplars into GRPO sampling and applies dual rewards (OmniReward and AI-Portrait) to improve photorealism, claiming better results than baselines on a new PortraitBench.
-
Z-Reward: Beyond Scalar Rewards by Internalizing Reasoning into Score Distributions
A teacher-student reward model learns reasoning-conditioned score distributions for text-to-image images, yielding ~89% preference accuracy and a 41% net human-preference gain when used for generator optimization.
-
Z-Reward: Beyond Scalar Rewards by Internalizing Reasoning into Score Distributions
Z-Reward trains a 27B reasoning teacher VLM on score distributions via GDSO and distills it via RISD into a 9B student, reaching 89.6% and 88.6% human preference accuracy with 41.3% optimization gain over SFT baseline.
-
TextAlign: Preference Alignment for Text Rendering with Hierarchical Rewards
TextAlign uses a hierarchical VLM reward for preference alignment to boost text accuracy in generative models like FLUX.1-dev.
-
TextAlign: Preference Alignment for Text Rendering with Hierarchical Rewards
TextAlign applies hierarchical VLM-based rewards to align text-to-image models for better glyph-level text rendering via GRPO and DPO.
-
Sketch Then Paint: Hierarchical Reinforcement Learning for Diffusion Multi-Modal Large Language Models
Proposes HT-GRPO with sketch-then-paint staged updates, prompt-conditioned importance ratios, and hierarchical credit assignment for dMLLMs, reporting gains on GenEval and DPG plus quality metrics.
-
Leveraging Verifier-Based Reinforcement Learning in Image Editing
Edit-R1 trains a CoT-based reasoning reward model with GCPO and uses it to boost image editing performance over VLMs and models like FLUX.1-kontext via GRPO.
-
Leveraging Verifier-Based Reinforcement Learning in Image Editing
Edit-R1 builds a CoT-based reasoning reward model (RRM) via SFT and GCPO, then applies it with GRPO to improve image editing models such as FLUX.1-kontext.
-
ViPO: Visual Preference Optimization at Scale
Poly-DPO improves robustness to noisy preference data in visual models, and the new ViPO dataset enables superior performance, with the method reducing to standard DPO on high-quality data.
-
Directly Aligning the Full Diffusion Trajectory with Fine-Grained Human Preference
Direct-Align and SRPO fine-tune FLUX using ground-truth-noise recovery and text-conditional relative rewards, improving human-evaluated realism and aesthetics roughly 3x.
-
Visko Orbis 1.0: A Live Model for Real-Time Interactive Long Video Generation
A chunk-wise streaming video model with bounded multi-scale memory and streaming 4K upscaling reports real-time interactive long-form generation and top Arena preference/stability scores.
-
MT-EditFlow: Reinforcement Learning for Multi-Turn Image Editing with Flow Matching
MT-EditFlow applies flow-matching RL with multi-reward aggregation to improve multi-turn image editing performance on models like FLUX.1-Kontext-dev by 6.85 points at turn-3.
-
FAIL: Flow Matching Adversarial Imitation Learning for Image Generation
Post-training of flow matching can be framed as adversarial imitation learning, and the proposed FAIL methods improve FLUX's generation quality using 13K expert images without preference pairs.
-
PSR: Scaling Multi-Subject Personalized Image Generation with Pairwise Subject-Consistency Rewards
A data-generation pipeline plus pairwise subject-consistency rewards in RL improve consistency and prompt adherence for multi-subject personalized image generation.
-
Principled RL for Flow Matching Emerges from the Chunk-level Policy Optimization
GCPO shifts RL policy optimization for flow matching from step-level to chunk-level grouping of consecutive denoising steps, reporting up to 43% relative gains over GRPO on T2I benchmarks and preference tasks.
-
RewardDance: Reward Scaling in Visual Generation
RewardDance reframes visual reward modeling as a yes/no judgment task in a VLM and reports consistent gains in text-to-image, text-to-video, and image-to-video generation as the reward model scales from 1B to 26B.
-
Training-Free Object-Background Compositional T2I via Dynamic Spatial Guidance and Multi-Path Pruning
A training-free method with time-dependent attention gating and trajectory pruning enhances object-background balance in diffusion-based image synthesis.
-
Follow-Your-Preference++: Rethinking Preference Alignment for Image Inpainting
Empirical study shows reward model ensembles mitigate biases like brightness and composition in preference data for image inpainting, yielding better performance than prior methods without architecture changes.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.