LMM-R1: Empowering 3B LMMs with Strong Reasoning Abilities Through Two-Stage Rule-Based RL

Enhancing reasoning in Large Multimodal Models (LMMs) faces unique challenges from the complex interplay between visual perception and logical reasoning, particularly in compact 3B-parameter architectures where architectural constraints limit reasoning capacity and modality alignment. While rule-based reinforcement learning (RL) excels in text-only domains, its multimodal extension confronts two critical barriers: (1) data limitations due to ambiguous answers and scarce complex reasoning examples, and (2) degraded foundational reasoning induced by multimodal pretraining. To address these challenges, we propose \textbf{LMM-R1}, a two-stage framework adapting rule-based RL for multimodal reasoning through \textbf{Foundational Reasoning Enhancement (FRE)} followed by \textbf{Multimodal Generalization Training (MGT)}. The FRE stage first strengthens reasoning abilities using text-only data with rule-based RL, then the MGT stage generalizes these reasoning capabilities to multimodal domains. Experiments on Qwen2.5-VL-Instruct-3B demonstrate that LMM-R1 achieves 4.83\% and 4.5\% average improvements over baselines in multimodal and text-only benchmarks, respectively, with a 3.63\% gain in complex Football Game tasks. These results validate that text-based reasoning enhancement enables effective multimodal generalization, offering a data-efficient paradigm that bypasses costly high-quality multimodal training data.

Label Distribution Learning with Biased Annotations by Learning Multi-Label Representation

Multi-label learning (MLL) has gained attention for its ability to represent real-world data. Label Distribution Learning (LDL), an extension of MLL to learning from label distributions, faces challenges in collecting accurate label distributions. To address the issue of biased annotations, based on the low-rank assumption, existing works recover true distributions from biased observations by exploring the label correlations. However, recent evidence shows that the label distribution tends to be full-rank, and naive apply of low-rank approximation on biased observation leads to inaccurate recovery and performance degradation. In this paper, we address the LDL with biased annotations problem from a novel perspective, where we first degenerate the soft label distribution into a hard multi-hot label and then recover the true label information for each instance. This idea stems from an insight that assigning hard multi-hot labels is often easier than assigning a soft label distribution, and it shows stronger immunity to noise disturbances, leading to smaller label bias. Moreover, assuming that the multi-label space for predicting label distributions is low-rank offers a more reasonable approach to capturing label correlations. Theoretical analysis and experiments confirm the effectiveness and robustness of our method on real-world datasets.

CILP-FGDI: Exploiting Vision-Language Model for Generalizable Person Re-Identification

The Visual Language Model, known for its robust cross-modal capabilities, has been extensively applied in various computer vision tasks. In this paper, we explore the use of CLIP (Contrastive Language-Image Pretraining), a vision-language model pretrained on large-scale image-text pairs to align visual and textual features, for acquiring fine-grained and domain-invariant representations in generalizable person re-identification. The adaptation of CLIP to the task presents two primary challenges: learning more fine-grained features to enhance discriminative ability, and learning more domain-invariant features to improve the model's generalization capabilities. To mitigate the first challenge thereby enhance the ability to learn fine-grained features, a three-stage strategy is proposed to boost the accuracy of text descriptions. Initially, the image encoder is trained to effectively adapt to person re-identification tasks. In the second stage, the features extracted by the image encoder are used to generate textual descriptions (i.e., prompts) for each image. Finally, the text encoder with the learned prompts is employed to guide the training of the final image encoder. To enhance the model's generalization capabilities to unseen domains, a bidirectional guiding method is introduced to learn domain-invariant image features. Specifically, domain-invariant and domain-relevant prompts are generated, and both positive (pulling together image features and domain-invariant prompts) and negative (pushing apart image features and domain-relevant prompts) views are used to train the image encoder. Collectively, these strategies contribute to the development of an innovative CLIP-based framework for learning fine-grained generalized features in person re-identification.

STHFL: Spatio-Temporal Heterogeneous Federated Learning

Federated learning is a new framework that protects data privacy and allows multiple devices to cooperate in training machine learning models. Previous studies have proposed multiple approaches to eliminate the challenges posed by non-iid data and inter-domain heterogeneity issues. However, they ignore the \textbf{spatio-temporal} heterogeneity formed by different data distributions of increasing task data in the intra-domain. Moreover, the global data is generally a long-tailed distribution rather than assuming the global data is balanced in practical applications. To tackle the \textbf{spatio-temporal} dilemma, we propose a novel setting named \textbf{Spatio-Temporal Heterogeneity} Federated Learning (STHFL). Specially, the Global-Local Dynamic Prototype (GLDP) framework is designed for STHFL. In GLDP, the model in each client contains personalized layers which can dynamically adapt to different data distributions. For long-tailed data distribution, global prototypes are served as complementary knowledge for the training on classes with few samples in clients without leaking privacy. As tasks increase in clients, the knowledge of local prototypes generated in previous tasks guides for training in the current task to solve catastrophic forgetting. Meanwhile, the global-local prototypes are updated through the moving average method after training local prototypes in clients. Finally, we evaluate the effectiveness of GLDP, which achieves remarkable results compared to state-of-the-art methods in STHFL scenarios.

BatStyler: Advancing Multi-category Style Generation for Source-free Domain Generalization

Source-Free Domain Generalization (SFDG) aims to develop a model that performs on unseen domains without relying on any source domains. However, the implementation remains constrained due to the unavailability of training data. Research on SFDG focus on knowledge transfer of multi-modal models and style synthesis based on joint space of multiple modalities, thus eliminating the dependency on source domain images. However, existing works primarily work for multi-domain and less-category configuration, but performance on multi-domain and multi-category configuration is relatively poor. In addition, the efficiency of style synthesis also deteriorates in multi-category scenarios. How to efficiently synthesize sufficiently diverse data and apply it to multi-category configuration is a direction with greater practical value. In this paper, we propose a method called BatStyler, which is utilized to improve the capability of style synthesis in multi-category scenarios. BatStyler consists of two modules: Coarse Semantic Generation and Uniform Style Generation modules. The Coarse Semantic Generation module extracts coarse-grained semantics to prevent the compression of space for style diversity learning in multi-category configuration, while the Uniform Style Generation module provides a template of styles that are uniformly distributed in space and implements parallel training. Extensive experiments demonstrate that our method exhibits comparable performance on less-category datasets, while surpassing state-of-the-art methods on multi-category datasets.

SoPo: Text-to-Motion Generation Using Semi-Online Preference Optimization

Text-to-motion generation is essential for advancing the creative industry but often presents challenges in producing consistent, realistic motions. To address this, we focus on fine-tuning text-to-motion models to consistently favor high-quality, human-preferred motions, a critical yet largely unexplored problem. In this work, we theoretically investigate the DPO under both online and offline settings, and reveal their respective limitation: overfitting in offline DPO, and biased sampling in online DPO. Building on our theoretical insights, we introduce Semi-online Preference Optimization (SoPo), a DPO-based method for training text-to-motion models using "semi-online" data pair, consisting of unpreferred motion from online distribution and preferred motion in offline datasets. This method leverages both online and offline DPO, allowing each to compensate for the other's limitations. Extensive experiments demonstrate that SoPo outperforms other preference alignment methods, with an MM-Dist of 3.25% (vs e.g. 0.76% of MoDiPO) on the MLD model, 2.91% (vs e.g. 0.66% of MoDiPO) on MDM model, respectively. Additionally, the MLD model fine-tuned by our SoPo surpasses the SoTA model in terms of R-precision and MM Dist. Visualization results also show the efficacy of our SoPo in preference alignment. Our project page is https://sopo-motion.github.io.

MageBench: Bridging Large Multimodal Models to Agents

LMMs have shown impressive visual understanding capabilities, with the potential to be applied in agents, which demand strong reasoning and planning abilities. Nevertheless, existing benchmarks mostly assess their reasoning abilities in language part, where the chain-of-thought is entirely composed of text.We consider the scenario where visual signals are continuously updated and required along the decision making process. Such vision-in-the-chain reasoning paradigm is more aligned with the needs of multimodal agents, while being rarely evaluated. In this paper, we introduce MageBench, a reasoning capability oriented multimodal agent benchmark that, while having light-weight environments, poses significant reasoning challenges and holds substantial practical value. This benchmark currently includes three types of environments: WebUI, Sokoban, and Football, comprising a total of 483 different scenarios. It thoroughly validates the agent's knowledge and engineering capabilities, visual intelligence, and interaction skills. The results show that only a few product-level models are better than random acting, and all of them are far inferior to human-level. More specifically, we found current models severely lack the ability to modify their planning based on visual feedback, as well as visual imagination, interleaved image-text long context handling, and other abilities. We hope that our work will provide optimization directions for LMM from the perspective of being an agent. We release our code and data at https://github.com/microsoft/MageBench.

Frequency-Guided Diffusion Model with Perturbation Training for Skeleton-Based Video Anomaly Detection

Video anomaly detection is an essential yet challenging open-set task in computer vision, often addressed by leveraging reconstruction as a proxy task. However, existing reconstruction-based methods encounter challenges in two main aspects: (1) limited model robustness for open-set scenarios, (2) and an overemphasis on, but restricted capacity for, detailed motion reconstruction. To this end, we propose a novel frequency-guided diffusion model with perturbation training, which enhances the model robustness by perturbation training and emphasizes the principal motion components guided by motion frequencies. Specifically, we first use a trainable generator to produce perturbative samples for perturbation training of the diffusion model. During the perturbation training phase, the model robustness is enhanced and the domain of the reconstructed model is broadened by training against this generator. Subsequently, perturbative samples are introduced for inference, which impacts the reconstruction of normal and abnormal motions differentially, thereby enhancing their separability. Considering that motion details originate from high-frequency information, we propose a masking method based on 2D discrete cosine transform to separate high-frequency information and low-frequency information. Guided by the high-frequency information from observed motion, the diffusion model can focus on generating low-frequency information, and thus reconstructing the motion accurately. Experimental results on five video anomaly detection datasets, including human-related and open-set benchmarks, demonstrate the effectiveness of the proposed method. Our code is available at https://github.com/Xiaofeng-Tan/FGDMAD-Code.

Redefining <Creative> in Dictionary: Towards a Enhanced Semantic Understanding of Creative Generation

Creativity, both in human and diffusion models, remains an inherently abstract concept; thus, simply adding "creative" to a prompt does not yield reliable semantic recognition by the model. In this work, we concretize the abstract notion of "creative" through the TP2O task, which aims to merge two unrelated concepts, and introduce CreTok, redefining "creative" as the token $\texttt{<CreTok>}$. This redefinition offers a more concrete and universally adaptable representation for concept blending. This redefinition occurs continuously, involving the repeated random sampling of text pairs with different concepts and optimizing cosine similarity between target and constant prompts. This approach enables $\texttt{<CreTok>}$ to learn a method for creative concept fusion. Extensive experiments demonstrate that the creative capability enabled by $\texttt{<CreTok>}$ substantially surpasses recent SOTA diffusion models and achieves superior creative generation. CreTok exhibits greater flexibility and reduced time overhead, as $\texttt{<CreTok>}$ can function as a universal token for any concept, facilitating creative generation without retraining.

Reduction-based Pseudo-label Generation for Instance-dependent Partial Label Learning

Instance-dependent Partial Label Learning (ID-PLL) aims to learn a multi-class predictive model given training instances annotated with candidate labels related to features, among which correct labels are hidden fixed but unknown. The previous works involve leveraging the identification capability of the training model itself to iteratively refine supervision information. However, these methods overlook a critical aspect of ID-PLL: the training model is prone to overfitting on incorrect candidate labels, thereby providing poor supervision information and creating a bottleneck in training. In this paper, we propose to leverage reduction-based pseudo-labels to alleviate the influence of incorrect candidate labels and train our predictive model to overcome this bottleneck. Specifically, reduction-based pseudo-labels are generated by performing weighted aggregation on the outputs of a multi-branch auxiliary model, with each branch trained in a label subspace that excludes certain labels. This approach ensures that each branch explicitly avoids the disturbance of the excluded labels, allowing the pseudo-labels provided for instances troubled by these excluded labels to benefit from the unaffected branches. Theoretically, we demonstrate that reduction-based pseudo-labels exhibit greater consistency with the Bayes optimal classifier compared to pseudo-labels directly generated from the predictive model.