Recent advancements in Large Language Models (LLMs) have shown that it is promising to utilize Process Reward Models (PRMs) as verifiers to enhance the performance of LLMs. However, current PRMs face three key challenges: (1) limited process supervision and generalization capabilities, (2) dependence on scalar value prediction without leveraging the generative abilities of LLMs, and (3) inability to scale the test-time compute of PRMs. In this work, we introduce GenPRM, a generative process reward model that performs explicit Chain-of-Thought (CoT) reasoning with code verification before providing judgment for each reasoning step. To obtain high-quality process supervision labels and rationale data, we propose Relative Progress Estimation (RPE) and a rationale synthesis framework that incorporates code verification. Experimental results on ProcessBench and several mathematical reasoning tasks show that GenPRM significantly outperforms prior PRMs with only 23K training data from MATH dataset. Through test-time scaling, a 1.5B GenPRM outperforms GPT-4o, and a 7B GenPRM surpasses Qwen2.5-Math-PRM-72B on ProcessBench. Additionally, GenPRM demonstrates strong abilities to serve as a critic model for policy model refinement. This work establishes a new paradigm for process supervision that bridges the gap between PRMs and critic models in LLMs. Our code, model, and data will be available in https://ryanliu112.github.io/GenPRM.

In recent years, the exceptional performance of generative models in generative tasks has sparked significant interest in their integration into decision-making processes. Due to their ability to handle complex data distributions and their strong model capacity, generative models can be effectively incorporated into decision-making systems by generating trajectories that guide agents toward high-reward state-action regions or intermediate sub-goals. This paper presents a comprehensive review of the application of generative models in decision-making tasks. We classify seven fundamental types of generative models: energy-based models, generative adversarial networks, variational autoencoders, normalizing flows, diffusion models, generative flow networks, and autoregressive models. Regarding their applications, we categorize their functions into three main roles: controllers, modelers and optimizers, and discuss how each role contributes to decision-making. Furthermore, we examine the deployment of these models across five critical real-world decision-making scenarios. Finally, we summarize the strengths and limitations of current approaches and propose three key directions for advancing next-generation generative directive models: high-performance algorithms, large-scale generalized decision-making models, and self-evolving and adaptive models.

Reward engineering is one of the key challenges in Reinforcement Learning (RL). Preference-based RL effectively addresses this issue by learning from human feedback. However, it is both time-consuming and expensive to collect human preference labels. In this paper, we propose a novel \textbf{V}ision-\textbf{L}anguage \textbf{P}reference learning framework, named \textbf{VLP}, which learns a vision-language preference model to provide preference feedback for embodied manipulation tasks. To achieve this, we define three types of language-conditioned preferences and construct a vision-language preference dataset, which contains versatile implicit preference orders without human annotations. The preference model learns to extract language-related features, and then serves as a preference annotator in various downstream tasks. The policy can be learned according to the annotated preferences via reward learning or direct policy optimization. Extensive empirical results on simulated embodied manipulation tasks demonstrate that our method provides accurate preferences and generalizes to unseen tasks and unseen language instructions, outperforming the baselines by a large margin.

Test-Time Scaling (TTS) is an important method for improving the performance of Large Language Models (LLMs) by using additional computation during the inference phase. However, current studies do not systematically analyze how policy models, Process Reward Models (PRMs), and problem difficulty influence TTS. This lack of analysis limits the understanding and practical use of TTS methods. In this paper, we focus on two core questions: (1) What is the optimal approach to scale test-time computation across different policy models, PRMs, and problem difficulty levels? (2) To what extent can extended computation improve the performance of LLMs on complex tasks, and can smaller language models outperform larger ones through this approach? Through comprehensive experiments on MATH-500 and challenging AIME24 tasks, we have the following observations: (1) The compute-optimal TTS strategy is highly dependent on the choice of policy model, PRM, and problem difficulty. (2) With our compute-optimal TTS strategy, extremely small policy models can outperform larger models. For example, a 1B LLM can exceed a 405B LLM on MATH-500. Moreover, on both MATH-500 and AIME24, a 0.5B LLM outperforms GPT-4o, a 3B LLM surpasses a 405B LLM, and a 7B LLM beats o1 and DeepSeek-R1, while with higher inference efficiency. These findings show the significance of adapting TTS strategies to the specific characteristics of each task and model and indicate that TTS is a promising approach for enhancing the reasoning abilities of LLMs.

Recent years have witnessed significant advancements in text-guided style transfer, primarily attributed to innovations in diffusion models. These models excel in conditional guidance, utilizing text or images to direct the sampling process. However, despite their capabilities, direct conditional guidance approaches often face challenges in balancing the expressiveness of textual semantics with the diversity of output results while capturing stylistic features. To address these challenges, we introduce ArtCrafter, a novel framework for text-to-image style transfer. Specifically, we introduce an attention-based style extraction module, meticulously engineered to capture the subtle stylistic elements within an image. This module features a multi-layer architecture that leverages the capabilities of perceiver attention mechanisms to integrate fine-grained information. Additionally, we present a novel text-image aligning augmentation component that adeptly balances control over both modalities, enabling the model to efficiently map image and text embeddings into a shared feature space. We achieve this through attention operations that enable smooth information flow between modalities. Lastly, we incorporate an explicit modulation that seamlessly blends multimodal enhanced embeddings with original embeddings through an embedding reframing design, empowering the model to generate diverse outputs. Extensive experiments demonstrate that ArtCrafter yields impressive results in visual stylization, exhibiting exceptional levels of stylistic intensity, controllability, and diversity.

Extracting physically plausible 3D human motion from videos is a critical task. Although existing simulation-based motion imitation methods can enhance the physical quality of daily motions estimated from monocular video capture, extending this capability to high-difficulty motions remains an open challenge. This can be attributed to some flawed motion clips in video-based motion capture results and the inherent complexity in modeling high-difficulty motions. Therefore, sensing the advantage of segmentation in localizing human body, we introduce a mask-based motion correction module (MCM) that leverages motion context and video mask to repair flawed motions, producing imitation-friendly motions; and propose a physics-based motion transfer module (PTM), which employs a pretrain and adapt approach for motion imitation, improving physical plausibility with the ability to handle in-the-wild and challenging motions. Our approach is designed as a plug-and-play module to physically refine the video motion capture results, including high-difficulty in-the-wild motions. Finally, to validate our approach, we collected a challenging in-the-wild test set to establish a benchmark, and our method has demonstrated effectiveness on both the new benchmark and existing public datasets.https://physicalmotionrestoration.github.io

Recent advances in Customized Concept Swapping (CCS) enable a text-to-image model to swap a concept in the source image with a customized target concept. However, the existing methods still face the challenges of inconsistency and inefficiency. They struggle to maintain consistency in both the foreground and background during concept swapping, especially when the shape difference is large between objects. Additionally, they either require time-consuming training processes or involve redundant calculations during inference. To tackle these issues, we introduce InstantSwap, a new CCS method that aims to handle sharp shape disparity at speed. Specifically, we first extract the bbox of the object in the source image automatically based on attention map analysis and leverage the bbox to achieve both foreground and background consistency. For background consistency, we remove the gradient outside the bbox during the swapping process so that the background is free from being modified. For foreground consistency, we employ a cross-attention mechanism to inject semantic information into both source and target concepts inside the box. This helps learn semantic-enhanced representations that encourage the swapping process to focus on the foreground objects. To improve swapping speed, we avoid computing gradients at each timestep but instead calculate them periodically to reduce the number of forward passes, which improves efficiency a lot with a little sacrifice on performance. Finally, we establish a benchmark dataset to facilitate comprehensive evaluation. Extensive evaluations demonstrate the superiority and versatility of InstantSwap. Project Page: https://instantswap.github.io/

We consider off-dynamics reinforcement learning (RL) where one needs to transfer policies across different domains with dynamics mismatch. Despite the focus on developing dynamics-aware algorithms, this field is hindered due to the lack of a standard benchmark. To bridge this gap, we introduce ODRL, the first benchmark tailored for evaluating off-dynamics RL methods. ODRL contains four experimental settings where the source and target domains can be either online or offline, and provides diverse tasks and a broad spectrum of dynamics shifts, making it a reliable platform to comprehensively evaluate the agent's adaptation ability to the target domain. Furthermore, ODRL includes recent off-dynamics RL algorithms in a unified framework and introduces some extra baselines for different settings, all implemented in a single-file manner. To unpack the true adaptation capability of existing methods, we conduct extensive benchmarking experiments, which show that no method has universal advantages across varied dynamics shifts. We hope this benchmark can serve as a cornerstone for future research endeavors.

We propose Lodge++, a choreography framework to generate high-quality, ultra-long, and vivid dances given the music and desired genre. To handle the challenges in computational efficiency, the learning of complex and vivid global choreography patterns, and the physical quality of local dance movements, Lodge++ adopts a two-stage strategy to produce dances from coarse to fine. In the first stage, a global choreography network is designed to generate coarse-grained dance primitives that capture complex global choreography patterns. In the second stage, guided by these dance primitives, a primitive-based dance diffusion model is proposed to further generate high-quality, long-sequence dances in parallel, faithfully adhering to the complex choreography patterns. Additionally, to improve the physical plausibility, Lodge++ employs a penetration guidance module to resolve character self-penetration, a foot refinement module to optimize foot-ground contact, and a multi-genre discriminator to maintain genre consistency throughout the dance. Lodge++ is validated by extensive experiments, which show that our method can rapidly generate ultra-long dances suitable for various dance genres, ensuring well-organized global choreography patterns and high-quality local motion.

Large Language Models (LLMs) have shown significant potential in designing reward functions for Reinforcement Learning (RL) tasks. However, obtaining high-quality reward code often involves human intervention, numerous LLM queries, or repetitive RL training. To address these issues, we propose CARD, a LLM-driven Reward Design framework that iteratively generates and improves reward function code. Specifically, CARD includes a Coder that generates and verifies the code, while a Evaluator provides dynamic feedback to guide the Coder in improving the code, eliminating the need for human feedback. In addition to process feedback and trajectory feedback, we introduce Trajectory Preference Evaluation (TPE), which evaluates the current reward function based on trajectory preferences. If the code fails the TPE, the Evaluator provides preference feedback, avoiding RL training at every iteration and making the reward function better aligned with the task objective. Empirical results on Meta-World and ManiSkill2 demonstrate that our method achieves an effective balance between task performance and token efficiency, outperforming or matching the baselines across all tasks. On 10 out of 12 tasks, CARD shows better or comparable performance to policies trained with expert-designed rewards, and our method even surpasses the oracle on 3 tasks.