arXiv subjects

Jianheng Liu

Publications and source records attributed to Jianheng Liu.

14 recordsLinked to original sources

arXiv2026

Darwin Mobile Agent: A Roadmap for Self-Evolution

The goal of artificial intelligence is to create agents capable of general, adaptive behaviour in open-ended environments. Guided by the "Bitter Lesson", we argue that the most effective path toward this goal is to systematically remove human priors and allow intelligence to naturally emerge through interaction with a "Big World" that is orders of magnitude more complex than the agent itself. We propose the mobile Graphical User Interface (GUI) as a practical proxy for such a world and introduce Darwin Mobile Agent, an open-source infrastructure designed as a foundation for autonomous reinforcement learning in this domain. This framework addresses the data-collection bottleneck in real-world mobile interactions by using an asynchronous agent-environment loop across parallel cloud-phone instances. We further propose a conceptual roadmap to systematically remove human priors from three fundamental pillars of a self-evolving agent: task curricula, outcome verification, and memory management. We validate that the Darwin infrastructure provides the stability and scalability required for the first stage of this roadmap: policy optimisation in the GUI domain. This work establishes the practical and theoretical foundation necessary to move toward truly autonomous, self-evolving GUI agents.

cs.AI↗

arXiv2026

Beyond Syntax: Action Semantics Learning for App Agents

The recent development of Large Language Models (LLMs) enables the rise of App agents that interpret user intent and operate smartphone Apps through actions such as clicking and scrolling. While prompt-based solutions with proprietary LLM APIs show promising ability, they incur heavy compute costs and external API dependency. Fine-tuning smaller open-source LLMs solves these limitations. However, current supervised fine-tuning methods use a syntax learning paradigm that forces agents to reproduce exactly the ground truth action strings, leading to out-of-distribution (OOD) vulnerability. To fill this gap, we propose Action Semantics Learning (ASL), a novel learning framework, where the learning objective is capturing the semantics of the ground truth actions. Specifically, inspired by the programming language theory, we define the action semantics for App agents as the state transition induced by the action in the user interface. Building on this insight, ASL employs a novel SEmantic Estimator~(SEE) to compute a semantic similarity to train the App agents in generating actions aligned with the semantics of ground truth actions, even when their syntactic forms differ. SEE is a flexible module that can be applied in both supervised and reinforcement fine-tuning paradigms. To support the effectiveness of ASL, we theoretically demonstrate the superior robustness of ASL for the OOD problem compared with the existing syntax learning paradigm. Extensive experiments across multiple offline and online benchmarks demonstrate that ASL significantly improves the accuracy and generalisation of App agents compared to existing methods.

cs.AI↗

arXiv2026

K^2-Agent: Co-Evolving Know-What and Know-How for Hierarchical Mobile Device Control

Existing mobile device control agents often perform poorly when solving complex tasks requiring long-horizon planning and precise operations, typically due to a lack of relevant task experience or unfamiliarity with skill execution. We propose K2-Agent, a hierarchical framework that models human-like cognition by separating and co-evolving declarative (knowing what) and procedural (knowing how) knowledge for planning and execution. K2-Agent's high level reasoner is bootstrapped from a single demonstration per task and runs a Summarize-Reflect-Locate-Revise (SRLR) loop to distill and iteratively refine task-level declarative knowledge through self-evolution. The low-level executor is trained with our curriculum-guided Group Relative Policy Optimization (C-GRPO), which (i) constructs a balanced sample pool using decoupled reward signals and (ii) employs dynamic demonstration injection to guide the model in autonomously generating successful trajectories for training. On the challenging AndroidWorld benchmark, K2-Agent achieves a 76.1% success rate using only raw screenshots and open-source backbones. Furthermore, K2-Agent shows powerful dual generalization: its high-level declarative knowledge transfers across diverse base models, while its low-level procedural skills achieve competitive performance on unseen tasks in ScreenSpot-v2 and Android-in-the-Wild (AitW).

cs.AI↗

arXiv2026

Efficient Swept Volume-Based Trajectory Generation for Arbitrary-Shaped Ground Robot Navigation

Navigating an arbitrary-shaped ground robot safely in cluttered environments remains a challenging problem. The existing trajectory planners that account for the robot's physical geometry severely suffer from the intractable runtime. To achieve both computational efficiency and Continuous Collision Avoidance (CCA) of arbitrary-shaped ground robot planning, we proposed a novel coarse-to-fine navigation framework that significantly accelerates planning. In the first stage, a sampling-based method selectively generates distinct topological paths that guarantee a minimum inflated margin. In the second stage, a geometry-aware front-end strategy is designed to discretize these topologies into full-state robot motion sequences while concurrently partitioning the paths into SE(2) sub-problems and simpler R2 sub-problems for back-end optimization. In the final stage, an SVSDF-based optimizer generates trajectories tailored to these sub-problems and seamlessly splices them into a continuous final motion plan. Extensive benchmark comparisons show that the proposed method is one to several orders of magnitude faster than the cutting-edge methods in runtime while maintaining a high planning success rate and ensuring CCA.

cs.RO↗

arXiv2025

Mesh-Learner: Texturing Mesh with Spherical Harmonics

In this paper, we present a 3D reconstruction and rendering framework termed Mesh-Learner that is natively compatible with traditional rasterization pipelines. It integrates mesh and spherical harmonic (SH) texture (i.e., texture filled with SH coefficients) into the learning process to learn each mesh s view-dependent radiance end-to-end. Images are rendered by interpolating surrounding SH Texels at each pixel s sampling point using a novel interpolation method. Conversely, gradients from each pixel are back-propagated to the related SH Texels in SH textures. Mesh-Learner exploits graphic features of rasterization pipeline (texture sampling, deferred rendering) to render, which makes Mesh-Learner naturally compatible with tools (e.g., Blender) and tasks (e.g., 3D reconstruction, scene rendering, reinforcement learning for robotics) that are based on rasterization pipelines. Our system can train vast, unlimited scenes because we transfer only the SH textures within the frustum to the GPU for training. At other times, the SH textures are stored in CPU RAM, which results in moderate GPU memory usage. The rendering results on interpolation and extrapolation sequences in the Replica and FAST-LIVO2 datasets achieve state-of-the-art performance compared to existing state-of-the-art methods (e.g., 3D Gaussian Splatting and M2-Mapping). To benefit the society, the code will be available at https://github.com/hku-mars/Mesh-Learner.

cs.CV↗

arXiv2025

GS-SDF: LiDAR-Augmented Gaussian Splatting and Neural SDF for Geometrically Consistent Rendering and Reconstruction

Digital twins are fundamental to the development of autonomous driving and embodied artificial intelligence. However, achieving high-granularity surface reconstruction and high-fidelity rendering remains a challenge. Gaussian splatting offers efficient photorealistic rendering but struggles with geometric inconsistencies due to fragmented primitives and sparse observational data in robotics applications. Existing regularization methods, which rely on render-derived constraints, often fail in complex environments. Moreover, effectively integrating sparse LiDAR data with Gaussian splatting remains challenging. We propose a unified LiDAR-visual system that synergizes Gaussian splatting with a neural signed distance field. The accurate LiDAR point clouds enable a trained neural signed distance field to offer a manifold geometry field. This motivates us to offer an SDF-based Gaussian initialization for physically grounded primitive placement and a comprehensive geometric regularization for geometrically consistent rendering and reconstruction. Experiments demonstrate superior reconstruction accuracy and rendering quality across diverse trajectories. To benefit the community, the codes are released at https://github.com/hku-mars/GS-SDF.

cs.RO↗

arXiv2025

Advancing Autonomous VLM Agents via Variational Subgoal-Conditioned Reinforcement Learning

State-of-the-art (SOTA) reinforcement learning (RL) methods have enabled vision-language model (VLM) agents to learn from interaction with online environments without human supervision. However, these methods often struggle with learning inefficiencies when applied to complex, real-world decision-making tasks with sparse rewards and long-horizon dependencies. We propose a novel framework, Variational Subgoal-Conditioned Reinforcement Learning (VSC-RL), advancing the VLM agents in resolving challenging decision-making tasks. Fundamentally distinct from existing methods, VSC-RL reformulates the decision-making problem as a variational subgoal-conditioned RL problem with the newly derived optimization objective, Subgoal Evidence Lower BOund (SGC-ELBO), which comprises two key components: (a) maximizing the subgoal-conditioned return, and (b) minimizing the divergence from a reference goal-conditioned policy. We theoretically and empirically demonstrate that the VSC-RL can efficiently improve the learning efficiency without compromising performance guarantees. Across a diverse set of challenging benchmarks, including mobile device and web control tasks, VSC-RL consistently outperforms existing SOTA methods, achieving superior learning efficiency and performance.

cs.LG↗

arXiv2025

DistRL: An Asynchronous Distributed Reinforcement Learning Framework for On-Device Control Agents

On-device control agents, especially on mobile devices, are responsible for operating mobile devices to fulfill users' requests, enabling seamless and intuitive interactions. Integrating Multimodal Large Language Models (MLLMs) into these agents enhances their ability to understand and execute complex commands, thereby improving user experience. However, fine-tuning MLLMs for on-device control presents significant challenges due to limited data availability and inefficient online training processes. This paper introduces DistRL, a novel framework designed to enhance the efficiency of online RL fine-tuning for mobile device control agents. DistRL employs centralized training and decentralized data acquisition to ensure efficient fine-tuning in the context of dynamic online interactions. Additionally, the framework is backed by our tailor-made RL algorithm, which effectively balances exploration with the prioritized utilization of collected data to ensure stable and robust training. Our experiments show that, on average, DistRL delivers a 3X improvement in training efficiency and enables training data collection 2.4X faster than the leading synchronous multi-machine methods. Notably, after training, DistRL achieves a 20% relative improvement in success rate compared to state-of-the-art methods on general Android tasks from an open benchmark, significantly outperforming existing approaches while maintaining the same training time. These results validate DistRL as a scalable and efficient solution, offering substantial improvements in both training efficiency and agent performance for real-world, in-the-wild device control tasks.

cs.LG↗

arXiv2025

OCMDP: Observation-Constrained Markov Decision Process

In many practical applications, decision-making processes must balance the costs of acquiring information with the benefits it provides. Traditional control systems often assume full observability, an unrealistic assumption when observations are expensive. We tackle the challenge of simultaneously learning observation and control strategies in such cost-sensitive environments by introducing the Observation-Constrained Markov Decision Process (OCMDP), where the policy influences the observability of the true state. To manage the complexity arising from the combined observation and control actions, we develop an iterative, model-free deep reinforcement learning algorithm that separates the sensing and control components of the policy. This decomposition enables efficient learning in the expanded action space by focusing on when and what to observe, as well as determining optimal control actions, without requiring knowledge of the environment's dynamics. We validate our approach on a simulated diagnostic task and a realistic healthcare environment using HeartPole. Given both scenarios, the experimental results demonstrate that our model achieves a substantial reduction in observation costs on average, significantly outperforming baseline methods by a notable margin in efficiency.

cs.AI↗

arXiv2024

Detect an Object At Once without Fine-tuning

When presented with one or a few photos of a previously unseen object, humans can instantly recognize it in different scenes. Although the human brain mechanism behind this phenomenon is still not fully understood, this work introduces a novel technical realization of this task. It consists of two phases: (1) generating a Similarity Density Map (SDM) by convolving the scene image with the given object image patch(es) so that the highlight areas in the SDM indicate the possible locations; (2) obtaining the object occupied areas in the scene through a Region Alignment Network (RAN). The RAN is constructed on a backbone of Deep Siamese Network (DSN), and different from the traditional DSNs, it aims to obtain the object accurate regions by regressing the location and area differences between the ground truths and the predicted ones indicated by the highlight areas in SDM. By pre-learning from labels annotated in traditional datasets, the SDM-RAN can detect previously unknown objects without fine-tuning. Experiments were conducted on the MS COCO, PASCAL VOC datasets. The results indicate that the proposed method outperforms state-of-the-art methods on the same task.

cs.CV↗

arXiv2024

Neural Surface Reconstruction and Rendering for LiDAR-Visual Systems

This paper presents a unified surface reconstruction and rendering framework for LiDAR-visual systems, integrating Neural Radiance Fields (NeRF) and Neural Distance Fields (NDF) to recover both appearance and structural information from posed images and point clouds. We address the structural visible gap between NeRF and NDF by utilizing a visible-aware occupancy map to classify space into the free, occupied, visible unknown, and background regions. This classification facilitates the recovery of a complete appearance and structure of the scene. We unify the training of the NDF and NeRF using a spatial-varying scale SDF-to-density transformation for levels of detail for both structure and appearance. The proposed method leverages the learned NDF for structure-aware NeRF training by an adaptive sphere tracing sampling strategy for accurate structure rendering. In return, NeRF further refines structural in recovering missing or fuzzy structures in the NDF. Extensive experiments demonstrate the superior quality and versatility of the proposed method across various scenarios. To benefit the community, the codes will be released at \url{https://github.com/hku-mars/M2Mapping}.

cs.RO↗

arXiv2024

Active Implicit Object Reconstruction using Uncertainty-guided Next-Best-View Optimization

Actively planning sensor views during object reconstruction is crucial for autonomous mobile robots. An effective method should be able to strike a balance between accuracy and efficiency. In this paper, we propose a seamless integration of the emerging implicit representation with the active reconstruction task. We build an implicit occupancy field as our geometry proxy. While training, the prior object bounding box is utilized as auxiliary information to generate clean and detailed reconstructions. To evaluate view uncertainty, we employ a sampling-based approach that directly extracts entropy from the reconstructed occupancy probability field as our measure of view information gain. This eliminates the need for additional uncertainty maps or learning. Unlike previous methods that compare view uncertainty within a finite set of candidates, we aim to find the next-best-view (NBV) on a continuous manifold. Leveraging the differentiability of the implicit representation, the NBV can be optimized directly by maximizing the view uncertainty using gradient descent. It significantly enhances the method's adaptability to different scenarios. Simulation and real-world experiments demonstrate that our approach effectively improves reconstruction accuracy and efficiency of view planning in active reconstruction tasks. The proposed system will open source at https://github.com/HITSZ-NRSL/ActiveImplicitRecon.git.

cs.RO↗

arXiv2024

Towards Large-Scale Incremental Dense Mapping using Robot-centric Implicit Neural Representation

Large-scale dense mapping is vital in robotics, digital twins, and virtual reality. Recently, implicit neural mapping has shown remarkable reconstruction quality. However, incremental large-scale mapping with implicit neural representations remains problematic due to low efficiency, limited video memory, and the catastrophic forgetting phenomenon. To counter these challenges, we introduce the Robot-centric Implicit Mapping (RIM) technique for large-scale incremental dense mapping. This method employs a hybrid representation, encoding shapes with implicit features via a multi-resolution voxel map and decoding signed distance fields through a shallow MLP. We advocate for a robot-centric local map to boost model training efficiency and curb the catastrophic forgetting issue. A decoupled scalable global map is further developed to archive learned features for reuse and maintain constant video memory consumption. Validation experiments demonstrate our method's exceptional quality, efficiency, and adaptability across diverse scales and scenes over advanced dense mapping methods using range sensors. Our system's code will be accessible at https://github.com/HITSZ-NRSL/RIM.git.

cs.RO↗

arXiv2023

RGB-D Inertial Odometry for a Resource-Restricted Robot in Dynamic Environments

Current simultaneous localization and mapping (SLAM) algorithms perform well in static environments but easily fail in dynamic environments. Recent works introduce deep learning-based semantic information to SLAM systems to reduce the influence of dynamic objects. However, it is still challenging to apply a robust localization in dynamic environments for resource-restricted robots. This paper proposes a real-time RGB-D inertial odometry system for resource-restricted robots in dynamic environments named Dynamic-VINS. Three main threads run in parallel: object detection, feature tracking, and state optimization. The proposed Dynamic-VINS combines object detection and depth information for dynamic feature recognition and achieves performance comparable to semantic segmentation. Dynamic-VINS adopts grid-based feature detection and proposes a fast and efficient method to extract high-quality FAST feature points. IMU is applied to predict motion for feature tracking and moving consistency check. The proposed method is evaluated on both public datasets and real-world applications and shows competitive localization accuracy and robustness in dynamic environments. Yet, to the best of our knowledge, it is the best-performance real-time RGB-D inertial odometry for resource-restricted platforms in dynamic environments for now. The proposed system is open source at: https://github.com/HITSZ-NRSL/Dynamic-VINS.git

cs.RO↗