Skywork R1V2: Multimodal Hybrid Reinforcement Learning for Reasoning

We present Skywork R1V2, a next-generation multimodal reasoning model and a major leap forward from its predecessor, Skywork R1V. At its core, R1V2 introduces a hybrid reinforcement learning paradigm that harmonizes reward-model guidance with rule-based strategies, thereby addressing the long-standing challenge of balancing sophisticated reasoning capabilities with broad generalization. To further enhance training efficiency, we propose the Selective Sample Buffer (SSB) mechanism, which effectively counters the ``Vanishing Advantages'' dilemma inherent in Group Relative Policy Optimization (GRPO) by prioritizing high-value samples throughout the optimization process. Notably, we observe that excessive reinforcement signals can induce visual hallucinations--a phenomenon we systematically monitor and mitigate through calibrated reward thresholds throughout the training process. Empirical results affirm the exceptional capability of R1V2, with benchmark-leading performances such as 62.6 on OlympiadBench, 79.0 on AIME2024, 63.6 on LiveCodeBench, and 74.0 on MMMU. These results underscore R1V2's superiority over existing open-source models and demonstrate significant progress in closing the performance gap with premier proprietary systems, including Gemini 2.5 and OpenAI o4-mini. The Skywork R1V2 model weights have been publicly released to promote openness and reproducibility https://huggingface.co/Skywork/Skywork-R1V2-38B.

Skywork R1V: Pioneering Multimodal Reasoning with Chain-of-Thought

We introduce Skywork R1V, a multimodal reasoning model extending the an R1-series Large language models (LLM) to visual modalities via an efficient multimodal transfer method. Leveraging a lightweight visual projector, Skywork R1V facilitates seamless multimodal adaptation without necessitating retraining of either the foundational language model or the vision encoder. To strengthen visual-text alignment, we propose a hybrid optimization strategy that combines Iterative Supervised Fine-Tuning (SFT) with Group Relative Policy Optimization (GRPO), significantly enhancing cross-modal integration efficiency. Additionally, we introduce an adaptive-length Chain-of-Thought distillation approach for reasoning data generation. This approach dynamically optimizes reasoning chain lengths, thereby enhancing inference efficiency and preventing excessive reasoning overthinking. Empirical evaluations demonstrate that Skywork R1V, with only 38B parameters, delivers competitive performance, achieving a score of 69.0 on the MMMU benchmark and 67.5 on MathVista. Meanwhile, it maintains robust textual reasoning performance, evidenced by impressive scores of 72.0 on AIME and 94.0 on MATH500. The Skywork R1V model weights have been publicly released to promote openness and reproducibility.

Neural Reflectance Fields for Radio-Frequency Ray Tracing

Ray tracing is widely employed to model the propagation of radio-frequency (RF) signal in complex environment. The modelling performance greatly depends on how accurately the target scene can be depicted, including the scene geometry and surface material properties. The advances in computer vision and LiDAR make scene geometry estimation increasingly accurate, but there still lacks scalable and efficient approaches to estimate the material reflectivity in real-world environment. In this work, we tackle this problem by learning the material reflectivity efficiently from the path loss of the RF signal from the transmitters to receivers. Specifically, we want the learned material reflection coefficients to minimize the gap between the predicted and measured powers of the receivers. We achieve this by translating the neural reflectance field from optics to RF domain by modelling both the amplitude and phase of RF signals to account for the multipath effects. We further propose a differentiable RF ray tracing framework that optimizes the neural reflectance field to match the signal strength measurements. We simulate a complex real-world environment for experiments and our simulation results show that the neural reflectance field can successfully learn the reflection coefficients for all incident angles. As a result, our approach achieves better accuracy in predicting the powers of receivers with significantly less training data compared to existing approaches.

LaNMP: A Language-Conditioned Mobile Manipulation Benchmark for Autonomous Robots

As robots that follow natural language become more capable and prevalent, we need a benchmark to holistically develop and evaluate their ability to solve long-horizon mobile manipulation tasks in large, diverse environments. To tackle this challenge, robots must use visual and language understanding, navigation, and manipulation capabilities. Existing datasets do not integrate all these aspects, restricting their efficacy as benchmarks. To address this gap, we present the Language, Navigation, Manipulation, Perception (LaNMP, pronounced Lamp) dataset and demonstrate the benefits of integrating these four capabilities and various modalities. LaNMP comprises 574 trajectories across eight simulated and real-world environments for long-horizon room-to-room pick-and-place tasks specified by natural language. Every trajectory consists of over 20 attributes, including RGB-D images, segmentations, and the poses of the robot body, end-effector, and grasped objects. We fine-tuned and tested two models in simulation, and evaluated a third on a physical robot, to demonstrate the benchmark's applicability in development and evaluation, as well as making models more sample efficient. The models performed suboptimally compared to humans; however, showed promise in increasing model sample efficiency, indicating significant room for developing more sample efficient multimodal mobile manipulation models using our benchmark.

NovelGym: A Flexible Ecosystem for Hybrid Planning and Learning Agents Designed for Open Worlds

As AI agents leave the lab and venture into the real world as autonomous vehicles, delivery robots, and cooking robots, it is increasingly necessary to design and comprehensively evaluate algorithms that tackle the ``open-world''. To this end, we introduce NovelGym, a flexible and adaptable ecosystem designed to simulate gridworld environments, serving as a robust platform for benchmarking reinforcement learning (RL) and hybrid planning and learning agents in open-world contexts. The modular architecture of NovelGym facilitates rapid creation and modification of task environments, including multi-agent scenarios, with multiple environment transformations, thus providing a dynamic testbed for researchers to develop open-world AI agents.

Dynamic Causal Disentanglement Model for Dialogue Emotion Detection

Emotion detection is a critical technology extensively employed in diverse fields. While the incorporation of commonsense knowledge has proven beneficial for existing emotion detection methods, dialogue-based emotion detection encounters numerous difficulties and challenges due to human agency and the variability of dialogue content.In dialogues, human emotions tend to accumulate in bursts. However, they are often implicitly expressed. This implies that many genuine emotions remain concealed within a plethora of unrelated words and dialogues.In this paper, we propose a Dynamic Causal Disentanglement Model based on hidden variable separation, which is founded on the separation of hidden variables. This model effectively decomposes the content of dialogues and investigates the temporal accumulation of emotions, thereby enabling more precise emotion recognition. First, we introduce a novel Causal Directed Acyclic Graph (DAG) to establish the correlation between hidden emotional information and other observed elements. Subsequently, our approach utilizes pre-extracted personal attributes and utterance topics as guiding factors for the distribution of hidden variables, aiming to separate irrelevant ones. Specifically, we propose a dynamic temporal disentanglement model to infer the propagation of utterances and hidden variables, enabling the accumulation of emotion-related information throughout the conversation. To guide this disentanglement process, we leverage the ChatGPT-4.0 and LSTM networks to extract utterance topics and personal attributes as observed information.Finally, we test our approach on two popular datasets in dialogue emotion detection and relevant experimental results verified the model's superiority.

Category Query Learning for Human-Object Interaction Classification

Unlike most previous HOI methods that focus on learning better human-object features, we propose a novel and complementary approach called category query learning. Such queries are explicitly associated to interaction categories, converted to image specific category representation via a transformer decoder, and learnt via an auxiliary image-level classification task. This idea is motivated by an earlier multi-label image classification method, but is for the first time applied for the challenging human-object interaction classification task. Our method is simple, general and effective. It is validated on three representative HOI baselines and achieves new state-of-the-art results on two benchmarks.

SOLQ: Segmenting Objects by Learning Queries

In this paper, we propose an end-to-end framework for instance segmentation. Based on the recently introduced DETR [1], our method, termed SOLQ, segments objects by learning unified queries. In SOLQ, each query represents one object and has multiple representations: class, location and mask. The object queries learned perform classification, box regression and mask encoding simultaneously in an unified vector form. During training phase, the mask vectors encoded are supervised by the compression coding of raw spatial masks. In inference time, mask vectors produced can be directly transformed to spatial masks by the inverse process of compression coding. Experimental results show that SOLQ can achieve state-of-the-art performance, surpassing most of existing approaches. Moreover, the joint learning of unified query representation can greatly improve the detection performance of original DETR. We hope our SOLQ can serve as a strong baseline for the Transformer-based instance segmentation. Code is available at https://github.com/megvii-research/SOLQ.

MOTR: End-to-End Multiple-Object Tracking with TRansformer

The key challenge in multiple-object tracking (MOT) task is temporal modeling of the object under track. Existing tracking-by-detection methods adopt simple heuristics, such as spatial or appearance similarity. Such methods, in spite of their commonality, are overly simple and insufficient to model complex variations, such as tracking through occlusion. Inherently, existing methods lack the ability to learn temporal variations from data. In this paper, we present MOTR, the first fully end-to-end multiple-object tracking framework. It learns to model the long-range temporal variation of the objects. It performs temporal association implicitly and avoids previous explicit heuristics. Built on Transformer and DETR, MOTR introduces the concept of "track query". Each track query models the entire track of an object. It is transferred and updated frame-by-frame to perform object detection and tracking, in a seamless manner. Temporal aggregation network combined with multi-frame training is proposed to model the long-range temporal relation. Experimental results show that MOTR achieves state-of-the-art performance. Code is available at https://github.com/megvii-model/MOTR.

Dynamic Metric Learning: Towards a Scalable Metric Space to Accommodate Multiple Semantic Scales

This paper introduces a new fundamental characteristic, \ie, the dynamic range, from real-world metric tools to deep visual recognition. In metrology, the dynamic range is a basic quality of a metric tool, indicating its flexibility to accommodate various scales. Larger dynamic range offers higher flexibility. In visual recognition, the multiple scale problem also exist. Different visual concepts may have different semantic scales. For example, ``Animal'' and ``Plants'' have a large semantic scale while ``Elk'' has a much smaller one. Under a small semantic scale, two different elks may look quite \emph{different} to each other . However, under a large semantic scale (\eg, animals and plants), these two elks should be measured as being \emph{similar}. %We argue that such flexibility is also important for deep metric learning, because different visual concepts indeed correspond to different semantic scales. Introducing the dynamic range to deep metric learning, we get a novel computer vision task, \ie, the Dynamic Metric Learning. It aims to learn a scalable metric space to accommodate visual concepts across multiple semantic scales. Based on three types of images, \emph{i.e.}, vehicle, animal and online products, we construct three datasets for Dynamic Metric Learning. We benchmark these datasets with popular deep metric learning methods and find Dynamic Metric Learning to be very challenging. The major difficulty lies in a conflict between different scales: the discriminative ability under a small scale usually compromises the discriminative ability under a large one, and vice versa. As a minor contribution, we propose Cross-Scale Learning (CSL) to alleviate such conflict. We show that CSL consistently improves the baseline on all the three datasets. The datasets and the code will be publicly available at https://github.com/SupetZYK/DynamicMetricLearning.