Zhaojian Li

Dong Chen, Xinda Qi, Yu Zheng et al.

Weed management plays an important role in many modern agricultural applications. Conventional weed control methods mainly rely on chemical herbicides or hand weeding, which are often cost-ineffective, environmentally unfriendly, or even posing a threat to food safety and human health. Recently, automated/robotic weeding using machine vision systems has seen increased research attention with its potential for precise and individualized weed treatment. However, dedicated, large-scale, and labeled weed image datasets are required to develop robust and effective weed identification systems but they are often difficult and expensive to obtain. To address this issue, data augmentation approaches, such as generative adversarial networks (GANs), have been explored to generate highly realistic images for agricultural applications. Yet, despite some progress, those approaches are often complicated to train or have difficulties preserving fine details in images. In this paper, we present the first work of applying diffusion probabilistic models (also known as diffusion models) to generate high-quality synthetic weed images based on transfer learning. Comprehensive experimental results show that the developed approach consistently outperforms several state-of-the-art GAN models, representing the best trade-off between sample fidelity and diversity and highest FID score on a common weed dataset, CottonWeedID15. In addition, the expanding dataset with synthetic weed images can apparently boost model performance on four deep learning (DL) models for the weed classification tasks. Furthermore, the DL models trained on CottonWeedID15 dataset with only 10% of real images and 90% of synthetic weed images achieve a testing accuracy of over 94%, showing high-quality of the generated weed samples. The codes of this study are made publicly available at https://github.com/DongChen06/DMWeeds.

Jiajia Li, Raju Thada Magar, Dong Chen et al.

Soybeans are a critical source of food, protein and oil, and thus have received extensive research aimed at enhancing their yield, refining cultivation practices, and advancing soybean breeding techniques. Within this context, soybean pod counting plays an essential role in understanding and optimizing production. Despite recent advancements, the development of a robust pod-counting algorithm capable of performing effectively in real-field conditions remains a significant challenge This paper presents a pioneering work of accurate soybean pod counting utilizing unmanned aerial vehicle (UAV) images captured from actual soybean fields in Michigan, USA. Specifically, this paper presents SoybeanNet, a novel point-based counting network that harnesses powerful transformer backbones for simultaneous soybean pod counting and localization with high accuracy. In addition, a new dataset of UAV-acquired images for soybean pod counting was created and open-sourced, consisting of 113 drone images with more than 260k manually annotated soybean pods captured under natural lighting conditions. Through comprehensive evaluations, SoybeanNet demonstrated superior performance over five state-of-the-art approaches when tested on the collected images. Remarkably, SoybeanNet achieved a counting accuracy of $84.51\%$ when tested on the testing dataset, attesting to its efficacy in real-world scenarios. The publication also provides both the source code (\url{https://github.com/JiajiaLi04/Soybean-Pod-Counting-from-UAV-Images}) and the labeled soybean dataset (\url{https://www.kaggle.com/datasets/jiajiali/uav-based-soybean-pod-images}), offering a valuable resource for future research endeavors in soybean pod counting and related fields.

Pengyu Chu, Zhaojian Li, Kaixiang Zhang et al.

Automated apple harvesting has attracted significant research interest in recent years due to its potential to revolutionize the apple industry, addressing the issues of shortage and high costs in labor. One key technology to fully enable efficient automated harvesting is accurate and robust apple detection, which is challenging due to complex orchard environments that involve varying lighting conditions and foliage/branch occlusions. Furthermore, clustered apples are common in the orchard, which brings additional challenges as the clustered apples may be identified as one apple. This will cause issues in localization for subsequent robotic operations. In this paper, we present the development of a novel deep learning-based apple detection framework, Occluder-Occludee Relational Network (O2RNet), for robust detection of apples in such clustered environments. This network exploits the occuluder-occludee relationship modeling head by introducing a feature expansion structure to enable the combination of layered traditional detectors to split clustered apples and foliage occlusions. More specifically, we collect a comprehensive apple orchard image dataset under different lighting conditions (overcast, front lighting, and back lighting) with frequent apple occlusions. We then develop a novel occlusion-aware network for apple detection, in which a feature expansion structure is incorporated into the convolutional neural networks to extract additional features generated by the original network for occluded apples. Comprehensive evaluations are performed, which show that the developed O2RNet outperforms state-of-the-art models with a higher accuracy of 94\% and a higher F1-score of 0.88 on apple detection.

Zhaojian Li, Tianshu Chu, Ilya V. Kolmanovsky et al.

There is a rapidly growing interest in the use of cloud computing for automotive vehicles to facilitate computation and data intensive tasks. Efficient utilization of on-demand cloud resources holds a significant potential to improve future vehicle safety, comfort, and fuel economy. In the meanwhile, issues like cyber security and resource allocation pose great challenges. In this paper, we treat the resource allocation problem for cloud-based automotive systems. Both private and public cloud paradigms are considered where a private cloud provides an internal, company-owned internet service dedicated to its own vehicles while a public cloud serves all subscribed vehicles. This paper establishes comprehensive models of cloud resource provisioning for both private and public cloud- based automotive systems. Complications such as stochastic communication delays and task deadlines are explicitly considered. In particular, a centralized resource provisioning model is developed for private cloud and chance constrained optimization is exploited to utilize the cloud resources for best Quality of Services. On the other hand, a decentralized auction-based model is developed for public cloud and reinforcement learning is employed to obtain an optimal bidding policy for a "selfish" agent. Numerical examples are presented to illustrate the effectiveness of the developed techniques.

Zhaojian Li, Firas Khasawneh, Xiang Yin et al.

Microscopic traffic models describe how cars interact with their neighbors in an uninterrupted traffic flow and are frequently used for reference in advanced vehicle control design. In this paper, we propose a novel mechanical system inspired microscopic traffic model using a mass-spring-damper-clutch system. This model naturally captures the ego vehicle's resistance to large relative speed and deviation from a (driver and speed dependent) desired relative distance when following the lead vehicle. Comparing to existing car following (CF) models, this model offers physically interpretable insights on the underlying CF dynamics, and is able to characterize the impact of the ego vehicle on the lead vehicle, which is neglected in existing CF models. Thanks to the nonlinear wave propagation analysis techniques for mechanical systems, the proposed model therefore has great scalability so that multiple mass-spring-damper-clutch system can be chained to study the macroscopic traffic flow. We investigate the stability of the proposed model on the system parameters and the time delay using spectral element method. We also develop a parallel recursive least square with inverse QR decomposition (PRLS-IQR) algorithm to identify the model parameters online. These real-time estimated parameters can be used to predict the driving trajectory that can be incorporated in advanced vehicle longitudinal control systems for improved safety and fuel efficiency. The PRLS-IQR is computationally efficient and numerically stable so it is suitable for online implementation. The traffic model and the parameter identification algorithm are validated on both simulations and naturalistic driving data from multiple drivers. Promising performance is demonstrated.

Zhaojian Li, Ilya Kolmanovsky, Ella Atkins et al.

This chapter presents an H-infinity filtering framework for cloud-aided semiactive suspension system with time-varying delays. In this system, road profile information is downloaded from a cloud database to facilitate onboard estimation of suspension states. Time-varying data transmission delays are considered and assumed to be bounded. A quarter-car linear suspension model is used and an H-infinity filter is designed with both onboard sensor measurements and delayed road profile information from the cloud. The filter design procedure is designed based on linear matrix inequalities (LMIs). Numerical simulation results are reported that illustrates the fusion of cloud-based and on-board information that can be achieved in Vehicleto- Cloud-to-Vehicle (V2C2V) implementation.

Jiajia Li, Mingle Xu, Lirong Xiang et al.

The past decade has witnessed the rapid development and adoption of ML & DL methodologies in agricultural systems, showcased by great successes in agricultural applications. However, these conventional ML/DL models have certain limitations: they heavily rely on large, costly-to-acquire labeled datasets for training, require specialized expertise for development and maintenance, and are mostly tailored for specific tasks, thus lacking generalizability. Recently, large pre-trained models, also known as FMs, have demonstrated remarkable successes in language, vision, and decision-making tasks across various domains. These models are trained on a large amount of data from multiple domains and modalities. Once trained, they can accomplish versatile tasks with just minor fine-tuning and minimal task-specific labeled data. Despite their proven effectiveness and huge potential, there has been little exploration of applying FMs to agriculture AI. Thus, this study aims to explore the potential of FMs in the field of smart agriculture. In particular, conceptual tools and technical background are presented to help the understanding of the problem space and uncover new research directions. To this end, recent FMs in the general CS domain are reviewed, and the models are categorized into four categories: language FMs, vision FMs, multimodal FMs, and reinforcement learning FMs. Then, the steps of developing agriculture FMs (AFMs) are outlined and potential applications in smart agriculture are discussed. Moreover, challenges and risks associated with developing AFMs are discussed, including model training, validation, and deployment. In summary, the advancement of AI in agriculture is explored by introducing AFMs as a promising paradigm that can significantly mitigate the reliance on extensive labeled datasets and enhance the efficiency, effectiveness, and generalization of agricultural AI systems.

Cheng Ouyang, Moeen Ul Islam, Dong Chen et al.

Soft robots offer significant advantages in safety and adaptability, yet achieving precise and dynamic control remains a major challenge due to their inherently complex and nonlinear dynamics. Recently, Data-enabled Predictive Control (DeePC) has emerged as a promising model-free approach that bypasses explicit system identification by directly leveraging input-output data. While DeePC has shown success in other domains, its application to soft robots remains underexplored, particularly for three-dimensional (3D) soft robotic systems. This paper addresses this gap by developing and experimentally validating an effective DeePC framework on a 3D, cable-driven soft arm. Specifically, we design and fabricate a soft robotic arm with a thick tubing backbone for stability, a dense silicone body with large cavities for strength and flexibility, and rigid endcaps for secure termination. Using this platform, we implement DeePC with singular value decomposition (SVD)-based dimension reduction for two key control tasks: fixed-point regulation and trajectory tracking in 3D space. Comparative experiments with a baseline model-based controller demonstrate DeePC's superior accuracy, robustness, and adaptability, highlighting its potential as a practical solution for dynamic control of soft robots.

Zhaojian Li, Bin Zhao, Yuan Yuan

Binaural stereo audio is recorded by imitating the way the human ear receives sound, which provides people with an immersive listening experience. Existing approaches leverage autoencoders and directly exploit visual spatial information to synthesize binaural stereo, resulting in a limited representation of visual guidance. For the first time, we propose a visually guided generative adversarial approach for generating binaural stereo audio from mono audio. Specifically, we develop a Stereo Audio Generation Model (SAGM), which utilizes shared spatio-temporal visual information to guide the generator and the discriminator to work separately. The shared visual information is updated alternately in the generative adversarial stage, allowing the generator and discriminator to deliver their respective guided knowledge while visually sharing. The proposed method learns bidirectional complementary visual information, which facilitates the expression of visual guidance in generation. In addition, spatial perception is a crucial attribute of binaural stereo audio, and thus the evaluation of stereo spatial perception is essential. However, previous metrics failed to measure the spatial perception of audio. To this end, a metric to measure the spatial perception of audio is proposed for the first time. The proposed metric is capable of measuring the magnitude and direction of spatial perception in the temporal dimension. Further, considering its function, it is feasible to utilize it instead of demanding user studies to some extent. The proposed method achieves state-of-the-art performance on 2 datasets and 5 evaluation metrics. Qualitative experiments and user studies demonstrate that the method generates space-realistic stereo audio.

Jiajia Li, Kyle Lammers, Xunyuan Yin et al.

Fruit harvesting poses a significant labor and financial burden for the industry, highlighting the critical need for advancements in robotic harvesting solutions. Machine vision-based fruit detection has been recognized as a crucial component for robust identification of fruits to guide robotic manipulation. Despite considerable progress in leveraging deep learning and machine learning techniques for fruit detection, a common shortfall is the inability to swiftly extend the developed models across different orchards and/or various fruit species. Additionally, the limited availability of pertinent data further compounds these challenges. In this work, we introduce MetaFruit, the largest publicly available multi-class fruit dataset, comprising 4,248 images and 248,015 manually labeled instances across diverse U.S. orchards. Furthermore, this study proposes an innovative open-set fruit detection system leveraging advanced Vision Foundation Models (VFMs) for fruit detection that can adeptly identify a wide array of fruit types under varying orchard conditions. This system not only demonstrates remarkable adaptability in learning from minimal data through few-shot learning but also shows the ability to interpret human instructions for subtle detection tasks. The performance of the developed foundation model is comprehensively evaluated using several metrics, which outperforms the existing state-of-the-art algorithms in both our MetaFruit dataset and other open-sourced fruit datasets, thereby setting a new benchmark in the field of agricultural technology and robotic harvesting. The MetaFruit dataset and detection framework are open-sourced to foster future research in vision-based fruit harvesting, marking a significant stride toward addressing the urgent needs of the agricultural sector.

Jiajia Li, Keyi Zhu, Qianwen Zhang et al.

Strawberries are among the most economically significant fruits in the United States, generating over $2 billion in annual farm-gate sales and accounting for approximately 13% of the total fruit production value. Plant phenotyping plays a vital role in selecting superior cultivars by characterizing plant traits such as morphology, canopy structure, and growth dynamics. However, traditional plant phenotyping methods are time-consuming, labor-intensive, and often destructive. Recently, neural rendering techniques, notably Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS), have emerged as powerful frameworks for high-fidelity 3D reconstruction. By capturing a sequence of multi-view images or videos around a target plant, these methods enable non-destructive reconstruction of complex plant architectures. Despite their promise, most current applications of 3DGS in agricultural domains reconstruct the entire scene, including background elements, which introduces noise, increases computational costs, and complicates downstream trait analysis. To address this limitation, we propose a novel object-centric 3D reconstruction framework incorporating a preprocessing pipeline that leverages the Segment Anything Model v2 (SAM-2) and alpha channel background masking to achieve clean strawberry plant reconstructions. This approach produces more accurate geometric representations while substantially reducing computational time. With a background-free reconstruction, our algorithm can automatically estimate important plant traits, such as plant height and canopy width, using DBSCAN clustering and Principal Component Analysis (PCA). Experimental results show that our method outperforms conventional pipelines in both accuracy and efficiency, offering a scalable and non-destructive solution for strawberry plant phenotyping.

Jiajia Li, Xinda Qi, Seyed Hamidreza Nabaei et al.

Plant phenotyping plays a pivotal role in understanding plant traits and their interactions with the environment, making it crucial for advancing precision agriculture and crop improvement. 3D reconstruction technologies have emerged as powerful tools for capturing detailed plant morphology and structure, offering significant potential for accurate and automated phenotyping. This paper provides a comprehensive review of the 3D reconstruction techniques for plant phenotyping, covering classical reconstruction methods, emerging Neural Radiance Fields (NeRF), and the novel 3D Gaussian Splatting (3DGS) approach. Classical methods, which often rely on high-resolution sensors, are widely adopted due to their simplicity and flexibility in representing plant structures. However, they face challenges such as data density, noise, and scalability. NeRF, a recent advancement, enables high-quality, photorealistic 3D reconstructions from sparse viewpoints, but its computational cost and applicability in outdoor environments remain areas of active research. The emerging 3DGS technique introduces a new paradigm in reconstructing plant structures by representing geometry through Gaussian primitives, offering potential benefits in both efficiency and scalability. We review the methodologies, applications, and performance of these approaches in plant phenotyping and discuss their respective strengths, limitations, and future prospects (https://github.com/JiajiaLi04/3D-Reconstruction-Plants). Through this review, we aim to provide insights into how these diverse 3D reconstruction techniques can be effectively leveraged for automated and high-throughput plant phenotyping, contributing to the next generation of agricultural technology.

Yujia Qin, Yining Ye, Junjie Fang et al.

This paper introduces UI-TARS, a native GUI agent model that solely perceives the screenshots as input and performs human-like interactions (e.g., keyboard and mouse operations). Unlike prevailing agent frameworks that depend on heavily wrapped commercial models (e.g., GPT-4o) with expert-crafted prompts and workflows, UI-TARS is an end-to-end model that outperforms these sophisticated frameworks. Experiments demonstrate its superior performance: UI-TARS achieves SOTA performance in 10+ GUI agent benchmarks evaluating perception, grounding, and GUI task execution. Notably, in the OSWorld benchmark, UI-TARS achieves scores of 24.6 with 50 steps and 22.7 with 15 steps, outperforming Claude (22.0 and 14.9 respectively). In AndroidWorld, UI-TARS achieves 46.6, surpassing GPT-4o (34.5). UI-TARS incorporates several key innovations: (1) Enhanced Perception: leveraging a large-scale dataset of GUI screenshots for context-aware understanding of UI elements and precise captioning; (2) Unified Action Modeling, which standardizes actions into a unified space across platforms and achieves precise grounding and interaction through large-scale action traces; (3) System-2 Reasoning, which incorporates deliberate reasoning into multi-step decision making, involving multiple reasoning patterns such as task decomposition, reflection thinking, milestone recognition, etc. (4) Iterative Training with Reflective Online Traces, which addresses the data bottleneck by automatically collecting, filtering, and reflectively refining new interaction traces on hundreds of virtual machines. Through iterative training and reflection tuning, UI-TARS continuously learns from its mistakes and adapts to unforeseen situations with minimal human intervention. We also analyze the evolution path of GUI agents to guide the further development of this domain.

Linjin He, Xinda Qi, Dong Chen et al.

Soft snake robots offer remarkable flexibility and adaptability in complex environments, yet their control remains challenging due to highly nonlinear dynamics. Existing model-based and bio-inspired controllers rely on simplified assumptions that limit performance. Deep reinforcement learning (DRL) has recently emerged as a promising alternative, but online training is often impractical because of costly and potentially damaging real-world interactions. Offline RL provides a safer option by leveraging pre-collected datasets, but it suffers from distribution shift, which degrades generalization to unseen scenarios. To overcome this challenge, we propose DiSA-IQL (Distribution-Shift-Aware Implicit Q-Learning), an extension of IQL that incorporates robustness modulation by penalizing unreliable state-action pairs to mitigate distribution shift. We evaluate DiSA-IQL on goal-reaching tasks across two settings: in-distribution and out-of-distribution evaluation. Simulation results show that DiSA-IQL consistently outperforms baseline models, including Behavior Cloning (BC), Conservative Q-Learning (CQL), and vanilla IQL, achieving higher success rates, smoother trajectories, and improved robustness. The codes are open-sourced to support reproducibility and to facilitate further research in offline RL for soft robot control.

Jiajia Li, Dong Chen, Xinda Qi et al.

The past decade has witnessed many great successes of machine learning (ML) and deep learning (DL) applications in agricultural systems, including weed control, plant disease diagnosis, agricultural robotics, and precision livestock management. Despite tremendous progresses, one downside of such ML/DL models is that they generally rely on large-scale labeled datasets for training, and the performance of such models is strongly influenced by the size and quality of available labeled data samples. In addition, collecting, processing, and labeling such large-scale datasets is extremely costly and time-consuming, partially due to the rising cost in human labor. Therefore, developing label-efficient ML/DL methods for agricultural applications has received significant interests among researchers and practitioners. In fact, there are more than 50 papers on developing and applying deep-learning-based label-efficient techniques to address various agricultural problems since 2016, which motivates the authors to provide a timely and comprehensive review of recent label-efficient ML/DL methods in agricultural applications. To this end, we first develop a principled taxonomy to organize these methods according to the degree of supervision, including weak supervision (i.e., active learning and semi-/weakly- supervised learning), and no supervision (i.e., un-/self- supervised learning), supplemented by representative state-of-the-art label-efficient ML/DL methods. In addition, a systematic review of various agricultural applications exploiting these label-efficient algorithms, such as precision agriculture, plant phenotyping, and postharvest quality assessment, is presented. Finally, we discuss the current problems and challenges, as well as future research directions. A well-classified paper list can be accessed at https://github.com/DongChen06/Label-efficient-in-Agriculture.

Haoming Wang, Haoyang Zou, Huatong Song et al. · pku

The development of autonomous agents for graphical user interfaces (GUIs) presents major challenges in artificial intelligence. While recent advances in native agent models have shown promise by unifying perception, reasoning, action, and memory through end-to-end learning, open problems remain in data scalability, multi-turn reinforcement learning (RL), the limitations of GUI-only operation, and environment stability. In this technical report, we present UI-TARS-2, a native GUI-centered agent model that addresses these challenges through a systematic training methodology: a data flywheel for scalable data generation, a stabilized multi-turn RL framework, a hybrid GUI environment that integrates file systems and terminals, and a unified sandbox platform for large-scale rollouts. Empirical evaluation demonstrates that UI-TARS-2 achieves significant improvements over its predecessor UI-TARS-1.5. On GUI benchmarks, it reaches 88.2 on Online-Mind2Web, 47.5 on OSWorld, 50.6 on WindowsAgentArena, and 73.3 on AndroidWorld, outperforming strong baselines such as Claude and OpenAI agents. In game environments, it attains a mean normalized score of 59.8 across a 15-game suite-roughly 60% of human-level performance-and remains competitive with frontier proprietary models (e.g., OpenAI o3) on LMGame-Bench. Additionally, the model can generalize to long-horizon information-seeking tasks and software engineering benchmarks, highlighting its robustness across diverse agent tasks. Detailed analyses of training dynamics further provide insights into achieving stability and efficiency in large-scale agent RL. These results underscore UI-TARS-2's potential to advance the state of GUI agents and exhibit strong generalization to real-world interactive scenarios.

Jiajia Li, Dong Chen, Xunyuan Yin et al.

Effective weed control plays a crucial role in optimizing crop yield and enhancing agricultural product quality. However, the reliance on herbicide application not only poses a critical threat to the environment but also promotes the emergence of resistant weeds. Fortunately, recent advances in precision weed management enabled by ML and DL provide a sustainable alternative. Despite great progress, existing algorithms are mainly developed based on supervised learning approaches, which typically demand large-scale datasets with manual-labeled annotations, which is time-consuming and labor-intensive. As such, label-efficient learning methods, especially semi-supervised learning, have gained increased attention in the broader domain of computer vision and have demonstrated promising performance. These methods aim to utilize a small number of labeled data samples along with a great number of unlabeled samples to develop high-performing models comparable to the supervised learning counterpart trained on a large amount of labeled data samples. In this study, we assess the effectiveness of a semi-supervised learning framework for multi-class weed detection, employing two well-known object detection frameworks, namely FCOS and Faster-RCNN. Specifically, we evaluate a generalized student-teacher framework with an improved pseudo-label generation module to produce reliable pseudo-labels for the unlabeled data. To enhance generalization, an ensemble student network is employed to facilitate the training process. Experimental results show that the proposed approach is able to achieve approximately 76\% and 96\% detection accuracy as the supervised methods with only 10\% of labeled data in CottenWeedDet3 and CottonWeedDet12, respectively. We offer access to the source code, contributing a valuable resource for ongoing semi-supervised learning research in weed detection and beyond.

Hongda Zhu, Yiwen Zhang, Bing Zhao et al.

Large Language Models (LLMs) have made significant strides in front-end code generation. However, existing benchmarks exhibit several critical limitations: many tasks are overly simplistic, test cases often lack rigor, and end-to-end validation is absent. These issues hinder the accurate assessment of model performance. To address these challenges, we present FrontendBench, a benchmark co-developed by humans and LLMs. FrontendBench categorizes tasks based on code functionality and incorporates interactive test scenarios, enabling a more comprehensive and practical evaluation of front-end code generation capabilities. The benchmark comprises 148 meticulously crafted prompt-test case pairs spanning five levels of web components, from basic UI elements to complex interactive features. Each task reflects realistic front-end development challenges. Furthermore, we introduce an automatic evaluation framework that executes generated code within a sandbox environment and assesses outcomes using predefined test scripts. This framework achieves a 90.54% agreement rate with expert human evaluations, demonstrating high reliability. We benchmark several state-of-the-art LLMs on FrontendBench and observe substantial performance disparities in handling real-world front-end tasks. These results highlight FrontendBench as a reliable and scalable benchmark, supporting consistent multimodal evaluation and providing a robust foundation for future research in front-end code generation. Our data and code will be released soon.

Keyi Zhu, Jiajia Li, Kaixiang Zhang et al.

Harvesting is a critical task in the tree fruit industry, demanding extensive manual labor and substantial costs, and exposing workers to potential hazards. Recent advances in automated harvesting offer a promising solution by enabling efficient, cost-effective, and ergonomic fruit picking within tight harvesting windows. However, existing harvesting technologies often indiscriminately harvest all visible and accessible fruits, including those that are unripe or undersized. This study introduces a novel foundation model-based framework for efficient apple ripeness and size estimation. Specifically, we curated two public RGBD-based Fuji apple image datasets, integrating expanded annotations for ripeness ("Ripe" vs. "Unripe") based on fruit color and image capture dates. The resulting comprehensive dataset, Fuji-Ripeness-Size Dataset, includes 4,027 images and 16,257 annotated apples with ripeness and size labels. Using Grounding-DINO, a language-model-based object detector, we achieved robust apple detection and ripeness classification, outperforming other state-of-the-art models. Additionally, we developed and evaluated six size estimation algorithms, selecting the one with the lowest error and variation for optimal performance. The Fuji-Ripeness-Size Dataset and the apple detection and size estimation algorithms are made publicly available, which provides valuable benchmarks for future studies in automated and selective harvesting.

Jingzhe Ding, Shengda Long, Changxin Pu et al.

Recent advances in coding agents suggest rapid progress toward autonomous software development, yet existing benchmarks fail to rigorously evaluate the long-horizon capabilities required to build complete software systems. Most prior evaluations focus on localized code generation, scaffolded completion, or short-term repair tasks, leaving open the question of whether agents can sustain coherent reasoning, planning, and execution over the extended horizons demanded by real-world repository construction. To address this gap, we present NL2Repo Bench, a benchmark explicitly designed to evaluate the long-horizon repository generation ability of coding agents. Given only a single natural-language requirements document and an empty workspace, agents must autonomously design the architecture, manage dependencies, implement multi-module logic, and produce a fully installable Python library. Our experiments across state-of-the-art open- and closed-source models reveal that long-horizon repository generation remains largely unsolved: even the strongest agents achieve below 40% average test pass rates and rarely complete an entire repository correctly. Detailed analysis uncovers fundamental long-horizon failure modes, including premature termination, loss of global coherence, fragile cross-file dependencies, and inadequate planning over hundreds of interaction steps. NL2Repo Bench establishes a rigorous, verifiable testbed for measuring sustained agentic competence and highlights long-horizon reasoning as a central bottleneck for the next generation of autonomous coding agents.

Musen Lin, Minghao Liu, Taoran Lu et al.

Graphical User Interface (GUI) Agents, powered by large language and vision-language models, hold promise for enabling end-to-end automation in digital environments. However, their progress is fundamentally constrained by the scarcity of scalable, high-quality trajectory data. Existing data collection strategies either rely on costly and inconsistent manual annotations or on synthetic generation methods that trade off between diversity and meaningful task coverage. To bridge this gap, we present GUI-ReWalk: a reasoning-enhanced, multi-stage framework for synthesizing realistic and diverse GUI trajectories. GUI-ReWalk begins with a stochastic exploration phase that emulates human trial-and-error behaviors, and progressively transitions into a reasoning-guided phase where inferred goals drive coherent and purposeful interactions. Moreover, it supports multi-stride task generation, enabling the construction of long-horizon workflows across multiple applications. By combining randomness for diversity with goal-aware reasoning for structure, GUI-ReWalk produces data that better reflects the intent-aware, adaptive nature of human-computer interaction. We further train Qwen2.5-VL-7B on the GUI-ReWalk dataset and evaluate it across multiple benchmarks, including Screenspot-Pro, OSWorld-G, UI-Vision, AndroidControl, and GUI-Odyssey. Results demonstrate that GUI-ReWalk enables superior coverage of diverse interaction flows, higher trajectory entropy, and more realistic user intent. These findings establish GUI-ReWalk as a scalable and data-efficient framework for advancing GUI agent research and enabling robust real-world automation.

Xinda Qi, Dong Chen, Zhaojian Li et al.

In this paper, we propose a novel technique, Back-stepping Experience Replay (BER), that is compatible with arbitrary off-policy reinforcement learning (RL) algorithms. BER aims to enhance learning efficiency in systems with approximate reversibility, reducing the need for complex reward shaping. The method constructs reversed trajectories using back-stepping transitions to reach random or fixed targets. Interpretable as a bi-directional approach, BER addresses inaccuracies in back-stepping transitions through a distillation of the replay experience during learning. Given the intricate nature of soft robots and their complex interactions with environments, we present an application of BER in a model-free RL approach for the locomotion and navigation of a soft snake robot, which is capable of serpentine motion enabled by anisotropic friction between the body and ground. In addition, a dynamic simulator is developed to assess the effectiveness and efficiency of the BER algorithm, in which the robot demonstrates successful learning (reaching a 100% success rate) and adeptly reaches random targets, achieving an average speed 48% faster than that of the best baseline approach.

Wei Zhou, Dong Chen, Jun Yan et al.

Autonomous driving has attracted significant research interests in the past two decades as it offers many potential benefits, including releasing drivers from exhausting driving and mitigating traffic congestion, among others. Despite promising progress, lane-changing remains a great challenge for autonomous vehicles (AV), especially in mixed and dynamic traffic scenarios. Recently, reinforcement learning (RL), a powerful data-driven control method, has been widely explored for lane-changing decision makings in AVs with encouraging results demonstrated. However, the majority of those studies are focused on a single-vehicle setting, and lane-changing in the context of multiple AVs coexisting with human-driven vehicles (HDVs) have received scarce attention. In this paper, we formulate the lane-changing decision making of multiple AVs in a mixed-traffic highway environment as a multi-agent reinforcement learning (MARL) problem, where each AV makes lane-changing decisions based on the motions of both neighboring AVs and HDVs. Specifically, a multi-agent advantage actor-critic network (MA2C) is developed with a novel local reward design and a parameter sharing scheme. In particular, a multi-objective reward function is proposed to incorporate fuel efficiency, driving comfort, and safety of autonomous driving. Comprehensive experimental results, conducted under three different traffic densities and various levels of human driver aggressiveness, show that our proposed MARL framework consistently outperforms several state-of-the-art benchmarks in terms of efficiency, safety and driver comfort.

Tianyu Shi, Dong Chen, Kaian Chen et al.

Reinforcement learning (RL) is a powerful data-driven control method that has been largely explored in autonomous driving tasks. However, conventional RL approaches learn control policies through trial-and-error interactions with the environment and therefore may cause disastrous consequences such as collisions when testing in real-world traffic. Offline RL has recently emerged as a promising framework to learn effective policies from previously-collected, static datasets without the requirement of active interactions, making it especially appealing for autonomous driving applications. Despite promising, existing offline RL algorithms such as Batch-Constrained deep Q-learning (BCQ) generally lead to rather conservative policies with limited exploration efficiency. To address such issues, this paper presents an enhanced BCQ algorithm by employing a learnable parameter noise scheme in the perturbation model to increase the diversity of observed actions. In addition, a Lyapunov-based safety enhancement strategy is incorporated to constrain the explorable state space within a safe region. Experimental results in highway and parking traffic scenarios show that our approach outperforms the conventional RL method, as well as state-of-the-art offline RL algorithms.

Kaixiang Zhang, Kyle Lammers, Pengyu Chu et al.

There is a growing need for robotic apple harvesting due to decreasing availability and rising cost in labor. Towards the goal of developing a viable robotic system for apple harvesting, this paper presents synergistic mechatronic design and motion control of a robotic apple harvesting prototype, which lays a critical foundation for future advancements. Specifically, we develop a deep learning-based fruit detection and localization system using an RGB-D camera. A three degree-of-freedom manipulator is then designed with a hybrid pneumatic/motor actuation mechanism to achieve fast and dexterous movements. A vacuum-based end-effector is used for apple detaching. These three components are integrated into a robotic apple harvesting prototype with simplicity, compactness, and robustness. Moreover, a nonlinear velocity-based control scheme is developed for the manipulator to achieve accurate and agile motion control. Test experiments are conducted to demonstrate the performance of the developed apple harvesting robot.

Pengyu Chu, Zhaojian Li, Kyle Lammers et al.

Robotic apple harvesting has received much research attention in the past few years due to growing shortage and rising cost in labor. One key enabling technology towards automated harvesting is accurate and robust apple detection, which poses great challenges as a result of the complex orchard environment that involves varying lighting conditions and foliage/branch occlusions. This letter reports on the development of a novel deep learning-based apple detection framework named DeepApple. Specifically, we first collect a comprehensive apple orchard dataset for 'Gala' and 'Blondee' apples, using a color camera, under different lighting conditions (sunny vs. overcast and front lighting vs. back lighting). We then develop a novel suppression Mask R-CNN for apple detection, in which a suppression branch is added to the standard Mask R-CNN to suppress non-apple features generated by the original network. Comprehensive evaluations are performed, which show that the developed suppression Mask R-CNN network outperforms state-of-the-art models with a higher F1-score of 0.905 and a detection time of 0.25 second per frame on a standard desktop computer.

Zhaonan Qu, Kaixiang Lin, Zhaojian Li et al.

Federated learning (FL) learns a model jointly from a set of participating devices without sharing each other's privately held data. The characteristics of non-i.i.d. data across the network, low device participation, high communication costs, and the mandate that data remain private bring challenges in understanding the convergence of FL algorithms, particularly regarding how convergence scales with the number of participating devices. In this paper, we focus on Federated Averaging (FedAvg), one of the most popular and effective FL algorithms in use today, as well as its Nesterov accelerated variant, and conduct a systematic study of how their convergence scale with the number of participating devices under non-i.i.d. data and partial participation in convex settings. We provide a unified analysis that establishes convergence guarantees for FedAvg under strongly convex, convex, and overparameterized strongly convex problems. We show that FedAvg enjoys linear speedup in each case, although with different convergence rates and communication efficiencies. For strongly convex and convex problems, we also characterize the corresponding convergence rates for the Nesterov accelerated FedAvg algorithm, which are the first linear speedup guarantees for momentum variants of FedAvg in convex settings. Empirical studies of the algorithms in various settings have supported our theoretical results.

Dong Chen, Longsheng Jiang, Yue Wang et al.

It is expected that many human drivers will still prefer to drive themselves even if the self-driving technologies are ready. Therefore, human-driven vehicles and autonomous vehicles (AVs) will coexist in a mixed traffic for a long time. To enable AVs to safely and efficiently maneuver in this mixed traffic, it is critical that the AVs can understand how humans cope with risks and make driving-related decisions. On the other hand, the driving environment is highly dynamic and ever-changing, and it is thus difficult to enumerate all the scenarios and hard-code the controllers. To face up these challenges, in this work, we incorporate a human decision-making model in reinforcement learning to control AVs for safe and efficient operations. Specifically, we adapt regret theory to describe a human driver's lane-changing behavior, and fit the personalized models to individual drivers for predicting their lane-changing decisions. The predicted decisions are incorporated in the safety constraints for reinforcement learning in training and in implementation. We then use an extended version of double deep Q-network (DDQN) to train our AV controller within the safety set. By doing so, the amount of collisions in training is reduced to zero, while the training accuracy is not impinged.

Tianshu Chu, Jie Wang, Lara Codecà et al.

Reinforcement learning (RL) is a promising data-driven approach for adaptive traffic signal control (ATSC) in complex urban traffic networks, and deep neural networks further enhance its learning power. However, centralized RL is infeasible for large-scale ATSC due to the extremely high dimension of the joint action space. Multi-agent RL (MARL) overcomes the scalability issue by distributing the global control to each local RL agent, but it introduces new challenges: now the environment becomes partially observable from the viewpoint of each local agent due to limited communication among agents. Most existing studies in MARL focus on designing efficient communication and coordination among traditional Q-learning agents. This paper presents, for the first time, a fully scalable and decentralized MARL algorithm for the state-of-the-art deep RL agent: advantage actor critic (A2C), within the context of ATSC. In particular, two methods are proposed to stabilize the learning procedure, by improving the observability and reducing the learning difficulty of each local agent. The proposed multi-agent A2C is compared against independent A2C and independent Q-learning algorithms, in both a large synthetic traffic grid and a large real-world traffic network of Monaco city, under simulated peak-hour traffic dynamics. Results demonstrate its optimality, robustness, and sample efficiency over other state-of-the-art decentralized MARL algorithms.