Cross-market recommendation (CMR) faces severe challenges from distribution shifts between data-rich source markets and sparse target markets. Existing methods rely on a pre-training and fine-tuning paradigm for knowledge transfer, yet suffer from two key limitations: i) the objective gap between pre-training and full-parameter fine-tuning causes loss of generalized knowledge from source markets; ii) the high computational costs of extensive fine-tuning hinder scalability. To this end, we propose DCMPT, a novel Distilled Cross-Market Prompt-Tuning approach. DCMPT reframes the problem under a more efficient pre-training and prompt-tuning paradigm. Instead of full fine-tuning, we adapt a pre-trained universal backbone by freezing its weights and injecting a minimal set of learnable prompts to form a "student" model. To effectively optimize these prompts on sparse data, we introduce a novel teacher-student architecture: a specialized "teacher" model, trained exclusively on the target market, provides dense, market-specific supervision. This guidance is delivered via a dual distillation strategy designed to transfer global ranking patterns and adapt to local consumer tastes. Extensive experiments on real-world market datasets demonstrate that DCMPT significantly outperforms state-of-the-art methods, achieving superior target market performance with substantial parameter-efficiency.

While printers are widely regarded as trusted peripherals, their internal mechanical execution reveals subtle vibrational patterns that can leak document structure. We present VibraPrint, a passive mmWave sensing system that infers high-level document attributes—such as page count, content density, and template type—as well as finer-grained structural cues including line count, per-line text amount, and average word-length trends. These properties emerge because layout-induced actuation patterns imprint low-frequency vibrations on the printer chassis, which are remotely captured using a 60 GHz radar without accessing content, print commands, or firmware. To extract meaningful structure from weak and heavily filtered signals, VibraPrint employs a two-stage recovery pipeline that combines global arc fitting with rhythm-aligned segment-wise refinement. Each segment is encoded using hybrid time–frequency features and processed by a structure-aware Transformer for multi-task inference. Evaluated on 500 print jobs across 20 printer models, VibraPrint achieves a mean page-count error of 1.05, over 90% accuracy for density and template prediction, and reliable estimation of per-line structure under distance and alignment variations. These results reveal a previously unrecognized class of structural side-channel leakage inherent to everyday printing workflows.

Recent advances in imitation learning for 3D robotic manipulation have shown promising results with diffusion-based policies. However, achieving human-level dexterity requires seamless integration of geometric precision and semantic understanding. We present G3Flow, a novel framework that constructs real-time semantic flow, a dynamic, object-centric 3D semantic representation by leveraging foundation models. Our approach uniquely combines 3D generative models for digital twin creation, vision foundation models for semantic feature extraction, and robust pose tracking for continuous semantic flow updates. This integration enables complete semantic understanding even under occlusions while eliminating manual annotation requirements. By incorporating semantic flow into diffusion policies, extensive experiments across five simulation tasks show that G3Flow consistently outperforms existing approaches, achieving up to 68.3% and 50.1% success rates on terminal-constrained manipulation and cross-object generalization respectively. Our results demonstrate the effectiveness of G3Flow in enhancing real-time dynamic semantic feature understanding for robotic policies.

Dexterous manipulation with contact-rich interactions is crucial for advanced robotics. While recent diffusion-based planning approaches show promise for simple manipulation tasks, they often produce unrealistic ghost states (e.g., the object automatically moves without hand contact) or lack adaptability when handling complex sequential interactions. In this work, we introduce DexHand-Diff, an interaction-aware diffusion planning framework for adaptive dexterous manipulation. DexHandDiff models joint state-action dynamics through a dual-phase diffusion process which consists of pre-interaction contact alignment and post-contact goal-directed control, enabling goal-adaptive generalizable dexterous manipulation. Additionally, we incorporate dynamics model-based dual guidance and leverage large language models for automated guidance function generation, enhancing generalizability for physical interactions and facilitating diverse goal adaptation through language cues. Experiments on physical interaction tasks such as door opening, pen and block reorientation, object relocation, and hammer striking demonstrate DexHandDiff’s effectiveness on goals outside training distributions, achieving over twice the average success rate (59.2% vs. 29.5%) compared to existing methods. Our framework achieves an average of 70.7% success rate on goal adaptive dexterous tasks, highlighting its robustness and flexibility in contact-rich manipulation.

In the rapidly advancing field of robotics, dual-arm co-ordination and complex object manipulation are essential capabilities for developing advanced autonomous systems. However, the scarcity of diverse, high-quality demonstration data and real-world-aligned evaluation benchmarks severely limits such development. To address this, we introduce RoboTwin, a generative digital twin framework that uses 3D generative foundation models and large language models to produce diverse expert datasets and provide a real-world-aligned evaluation platform for dual-arm robotic tasks. Specifically, RoboTwin creates varied digital twins of objects from single 2D images, generating realistic and interactive scenarios. It also introduces a spatial relation-aware code generation framework that combines object annotations with large language models to break down tasks, determine spatial constraints, and generate precise robotic movement code. Our framework offers a comprehensive benchmark with both simulated and real-world data, enabling standardized evaluation and better alignment between simulated training and real-world performance. We validated our approach using the open-source COBOT Magic Robot platform. Policies pre-trained on RoboTwin-generated data and fine-tuned with limited real-world samples demonstrate significant potential for enhancing dual-arm robotic manipulation systems by improving success rates by over 70% for single-arm tasks and over 40% for dual-arm tasks compared to models trained solely on real-world data.

With the rapid development of Large Vision Language Models, the focus of Graphical User Interface (GUI) agent tasks shifts from single-screen tasks to complex screen navigation challenges. However, real-world GUI environments, such as PC software and mobile Apps, are often complex and proprietary, making it difficult to obtain the comprehensive environment information needed for agent training and evaluation. This limitation hinders systematic investigation and benchmarking of agent navigation capabilities. To address this limitation, we introduce GUI Exploration Lab, a simulation environment engine for GUI agent navigation research that enables flexible definition and composition of screens, icons, and navigation graphs, while providing full access to environment information for comprehensive agent training and evaluation. Through extensive experiments, we find that supervised fine-tuning enables effective memorization of fundamental knowledge, serving as a crucial foundation for subsequent training. Building on this, single-turn reinforcement learning further enhances generalization to unseen scenarios. Finally, multi-turn reinforcement learning encourages the development of exploration strategies through interactive trial and error, leading to further improvements in screen navigation performance. We validate our methods on both static and interactive benchmarks, demonstrating that our findings generalize effectively to real-world scenarios. These findings demonstrate the advantages of reinforcement learning approaches in GUI navigation and offer practical guidance for building more capable and generalizable GUI agents.

Collaborative edge computing (CEC) has emerged as a promising paradigm, enabling edge nodes to collaborate and execute tasks from end devices. Task offloading is a fundamental problem in CEC that decides when and where tasks are executed upon the arrival of tasks. However, the mobility of users often results in unstable connections, leading to network failures and resource underutilization. Existing works have not adequately addressed joint mobility-aware dependent task offloading and network flow scheduling, resulting in network congestion and suboptimal performance. To address this, we formulate an online joint mobility-aware dependent task offloading and bandwidth allocation problem, to improve the quality of service by reducing task completion time and energy consumption. We introduce a Mobility-aware Digital Twin-assisted Deep Reinforcement Learning (MDT-DRL) algorithm. Our digital twin model equips the reinforcement learning process by providing future states of mobile users, enabling efficient offloading plans for adapting to the mobile CEC system. Experimental results on real-world and synthetic datasets show that MDT-DRL surpasses state-of-the-art baselines on average task completion time and energy consumption.

Diffusion models have demonstrated strong potential for robotic trajectory planning. However, generating coherent trajectories from high-level instructions remains challenging, especially for long-range composition tasks requiring multiple sequential skills. We propose SkillDiffuser, an end-to-end hierarchical planning framework integrating interpretable skill learning with conditional diffusion planning to address this problem. At the higher level, the skill abstraction module learns discrete, human-understandable skill representations from visual observations and language instructions. These learned skill embeddings are then used to condition the diffusion model to generate customized latent trajectories aligned with the skills. This allows generating diverse state trajectories that adhere to the learnable skills. By integrating skill learning with conditional trajectory generation, SkillDiffuser produces coherent behavior following abstract instructions across diverse tasks. Experiments on multitask robotic manipulation benchmarks like Meta-World and LOReL demonstrate state-of-the-art performance and human-interpretable skill representations from SkillDiffuser. More visualization results and information could be found on our website.

Robotic behavior synthesis, the problem of understanding multimodal inputs and generating precise physical control for robots, is an important part of Embodied AI. Despite successes in applying multimodal large language models for high-level understanding, it remains challenging to translate these conceptual understandings into detailed robotic actions while achieving generalization across various scenarios. In this paper, we propose a tree-structured multimodal code generation framework for generalized robotic behavior synthesis, termed RoboCodeX. RoboCodeX decomposes high-level human instructions into multiple object-centric manipulation units consisting of physical preferences such as affordance and safety constraints, and applies code generation to introduce generalization ability across various robotics platforms. To further enhance the capability to map conceptual and perceptual understanding into control commands, a specialized multimodal reasoning dataset is collected for pre-training and an iterative self-updating methodology is introduced for supervised fine-tuning. Extensive experiments demonstrate that RoboCodeX achieves state-of-the-art performance in both simulators and real robots on four different kinds of manipulation tasks and one embodied navigation task.

Multi-robot reinforcement learning (MRRL) is a promising approach to solving cooperation problems and has been widely adopted in many applications. In the past decades, researchers have proposed various approaches to improve the efficiency of MRRL. However, most of them are trained and evaluated only in simulated environments with simple interaction scenarios. The problem of how these methods perform in the real-world environment with complex interaction scenarios remains unsolved. To meet this emergent need, we introduce a scalable multi-robot reinforcement learning platform (SMART) for training and evaluation. Specifically, SMART consists of two components: 1) a simulation environment with an uncertainty-aware social agent model that provides a variety of complex interaction scenarios for training and 2) a real-world multi-robot system for realistic performance evaluation. To evaluate the generalizability of MRRL baselines, we introduce a novel generalization metric that takes into account their performance across changes in the environment as well as the policies of other agents. Furthermore, we conduct a case study on the multi-vehicle cooperative lane change and summarize the unique challenges of MRRL, which are rarely considered previously. Finally, we open-source the simulation environments, associated benchmark tasks, and state-of-the-art baselines to encourage and empower MRRL research. Our code is available at https://github.com/Blackmamba-xuan/MRST.

Collaborative edge computing (CEC) is an emerging computing paradigm in which edge nodes collaborate to perform tasks from end devices. Task offloading decides when and at which edge node tasks are executed. Most existing studies assume task profiles and network conditions are known in advance, which can hardly adapt to dynamic real-world computation environments. Some learning-based methods use online task offloading without considering task dependency and network flow scheduling, leading to underutilized resources and flow congestion. We study Online Dependent Task Offloading (ODTO) in CEC, jointly optimizing network flow scheduling to optimize quality of service by reducing task completion time and energy consumption. The challenge of ODTO lies in how to offload dependent tasks and schedule network flows in dynamic networks. We model ODTO as the Markov Decision Process (MDP) and propose an Asynchronous Deep Progressive Reinforcement Learning (ADPRL) approach that optimizes offloading and bandwidth decisions. We design a novel dependency-aware reward mechanism to address task dependency and dynamic networks. Extensive experiments on the Alibaba cluster trace dataset and synthetic dataset indicate that our algorithm outperforms heuristic and learning-based methods in average task completion time and energy consumption.

Diffusion models have demonstrated their powerful generative capability in many tasks, with great potential to serve as a paradigm for offline reinforcement learning. However, the quality of the diffusion model is limited by the insufficient diversity of training data, which hinders the performance of planning and the generalizability to new tasks. This paper introduces AdaptDiffuser, an evolutionary planning method with diffusion that can self-evolve to improve the diffusion model hence a better planner, not only for seen tasks but can also adapt to unseen tasks. AdaptDiffuser enables the generation of rich synthetic expert data for goal-conditioned tasks using guidance from reward gradients. It then selects high-quality data via a discriminator to finetune the diffusion model, which improves the generalization ability to unseen tasks. Empirical experiments on two benchmark environments and two carefully designed unseen tasks in KUKA industrial robot arm and Maze2D environments demonstrate the effectiveness of AdaptDiffuser. For example, AdaptDiffuser not only outperforms the previous art Diffuser by 20.8% on Maze2D and 7.5% on MuJoCo locomotion, but also adapts better to new tasks, e.g., KUKA pick-and-place, by 27.9% without requiring additional expert data. More visualization results and demo videos could be found on our project page.

Recently, diffusion model shines as a promising backbone for the sequence modeling paradigm in offline reinforcement learning(RL). However, these works mostly lack the generalization ability across tasks with reward or dynamics change. To tackle this challenge, in this paper we propose a task-oriented conditioned diffusion planner for offline meta-RL(MetaDiffuser), which considers the generalization problem as conditional trajectory generation task with contextual representation. The key is to learn a context conditioned diffusion model which can generate task-oriented trajectories for planning across diverse tasks. To enhance the dynamics consistency of the generated trajectories while encouraging trajectories to achieve high returns, we further design a dual-guided module in the sampling process of the diffusion model. The proposed framework enjoys the robustness to the quality of collected warm-start data from the testing task and the flexibility to incorporate with different task representation method. The experiment results on MuJoCo benchmarks show that MetaDiffuser outperforms other strong offline meta-RL baselines, demonstrating the outstanding conditional generation ability of diffusion architecture.

Pipelines have become one of the most important infrastructures in the city. Over time, they are prone to aging, cracks, corrosion, and the demand for regular inspection is gradually increasing. Robotic solutions are effective methods for in-pipe inspection. However, existing In-pipe Inspection Robots (IPIR) require that the inner diameter of the pipe is fixed in the application scenarios, and need extra labor to control the robot and handle the cable. In this work, we design and develop a deformable robot to adapt to pipes with different inner diameters. Specifically, the passive elastic hinge is used by us to make the robot fully in contact with the pipe, generating enough friction to ensure that the robot is attached to the inner wall of the pipe. An edge device is deployed on the robot, generating velocity commands of wheels through the data from Inertial Measurement Unit (IMU), which eliminates the need for external devices. Experimental results demonstrate that the robot can move in horizontal and vertical pipelines, as well as traverse through pipe joints and scenarios where there is dirty or small obstacle.

Many real-world applications can be formulated as multi-agent cooperation problems, such as network packet routing and coordination of autonomous vehicles. The emergence of deep reinforcement learning (DRL) provides a promising approach for multi-agent cooperation through the interaction of the agents and environments. However, traditional DRL solutions suffer from the high dimensions of multiple agents with continuous action space during policy search. Besides, the dynamicity of agents’ policies makes the training non-stationary. To tackle the issues, we propose a hierarchical reinforcement learning approach with high-level decision-making and low-level individual control for efficient policy search. In particular, the cooperation of multiple agents can be learned in high-level discrete action space efficiently. At the same time, the low-level individual control can be reduced to single-agent reinforcement learning. In addition to hierarchical reinforcement learning, we propose an opponent modeling network to model other agents’ policies during the learning process. In contrast to end-to-end DRL approaches, our approach reduces the learning complexity by decomposing the overall task into sub-tasks in a hierarchical way. To evaluate the efficiency of our approach, we conduct a real-world case study in the cooperative lane change scenario. Both simulation and real-world experiments show the superiority of our approach in the collision rate and convergence speed.

Summary Multi‐robot systems are widely used to handle complex and cooperative missions in various industrial applications. Although robotic middleware has become the key to reducing the complexity of multi‐robot application development, existing works still have limitations in controlling multiple robots to perform missions cooperatively. To enable multi‐robot cooperation, middleware should provide high‐level abstraction support, dynamic configuration, communication, and synchronization. In this article, we proposeGraphWare, a novel middleware that provides a graph‐based programming abstraction and its underlying runtime kernel for programming and building multi‐robot cooperation applications. The graph‐based programming abstraction can express cooperative missions without exposing the complexity of managing multiple robots. The runtime kernel configures and manages multiple heterogeneous robots to intelligently perform cooperative missions. We implementGraphWareand evaluate its performance with ball collection missions which are cooperatively accomplished by a group of mobile robots, and study the fault‐tolerance, flexibility, and scalability of the middleware in the realistic simulation. The experimental results demonstrate thatGraphWarefacilitates the multi‐robot cooperative mission with efficient mission completion time, high success rate, and marginal runtime overhead.

Data usually resides on a manifold, and the minimal dimension of such a manifold is called its intrinsic dimension. This fundamental data property is not considered in the generative adversarial network (GAN) model along with its its variants; such that original data and generated data often hold different intrinsic dimensions. The different intrinsic dimensions of both generated and original data may cause generated data distribution to not match original data distribution completely, and it certainly will hurt the quality of generated data. In this study, we first show that GAN is often unable to generate simulation data, holding the same intrinsic dimension as the original data with both theoretical analysis and experimental illustration. Next, we propose a new model, called Hausdorff GAN, which removes the issue of different intrinsic dimensions and introduces the Hausdorff metric into GAN training to generate higher quality data. This provides new insights into the success of Hausdorff GAN. Specifically, we utilize a mapping function to map both original and generated data into the same manifold. We then calculate the Hausdorff distance to measure the difference between the mapped original data and the mapped generated data, toward pushing generated data to the side of original data. Finally, we conduct extensive experiments (using MNIST, CIFAR10, and CelebA datasets) to demonstrate the significant performance improvement of the Hausdorff GAN in achieving the largest Inception Score and the smallest Frechet inception distance (FID) score as well as producing diverse generated data at different resolutions.

Federated Learning (FL) creates an ecosystem for multiple agents to collaborate on building models with data privacy consideration. The method for contribution measurement of each agent in the FL system is critical for fair credits allocation but few are proposed. In this paper, we develop a real-time contribution measurement method FedCM that is simple but powerful. The method defines the impact of each agent, comprehensively considers the current round and the previous round to obtain the contribution rate of each agent with attention aggregation. Moreover, FedCM updates contribution every round, which enable it to perform in real-time. Real-time is not considered by the existing approaches, but it is critical for FL systems to allocate computing power, communication resources, etc. Compared to the state-of-the-art method, the experimental results show that FedCM is more sensitive to data quantity and data quality under the premise of real-time. Furthermore, we developed federated learning open-source software based on FedCM. The software has been applied to identify COVID-19 based on medical images.

Multiview representations reveal the latent information of the data from different perspectives, consistency and complementarity. Unlike most multiview learning approaches, which focus only one perspective, in this paper, we propose a novel unsupervised multiview learning algorithm, called comprehensive latent factor learning (CLFL), which jointly exploits both consistent and complementary information among multiple views. CLFL adopts a non-negative matrix factorization based formulation to learn the latent factors. It learns the weights of different views automatically which makes the representation more accurate. Experiment results on a synthetic and several real datasets demonstrate the effectiveness of our approach.

Recently, much attention is paid to multi-robot systems due to their widespread applications such as warehouse robotics, persistent surveillance, and exploration of unknown environments. Although urgently required by the applications, coordination among multiple robots remains to be challenging. Among the problems of multi-robot coordination, pattern formation serves a fundamental one. It aims to control a group of robots to form a desired shape with some certain goals such as best formation quality, minimum makespan or minimum total distance. Existing works mainly focus on the formation of certain patterns, such as repeating squares or a circle. those approaches cannot be generalized to arbitrary pattern formation. In this paper, we propose a decentralized algorithm for a multi-robot system to generate a given formation with an arbitrary repeating pattern. We introduce basic pattern graph and assembling graph to define a repeating pattern and formation quality for measurement. Towards solving the repeating pattern formation problem, our approach is divided into two phases. The robots are grouped into multiple basic patterns in the first phase, and the patterns are assembled level by level in the second phase. Simulations and real-world experiments indicate the effectiveness and practicability of our approach.