arXiv subjects

Duo Wu

Publications and source records attributed to Duo Wu.

11 recordsLinked to original sources

arXiv2026

Learning with Challenges: Adaptive Difficulty-Aware Data Generation for Mobile GUI Agent Training

Large-scale, high-quality interaction trajectories are essential for advancing mobile Graphical User Interface (GUI) agents. While existing methods typically rely on labor-intensive human demonstrations or automated model exploration to generate GUI trajectories, they lack fine-grained control over task difficulty. This fundamentally restricts learning effectiveness due to the mismatch between the training difficulty and the agent's capabilities. Inspired by how humans acquire skills through progressively challenging tasks, we propose MobileGen, a novel data generation framework that adaptively aligns training difficulty with the GUI agent's capability frontier. Specifically, MobileGen explicitly decouples task difficulty into structural (e.g., trajectory length) and semantic (e.g., task goal) dimensions. It then iteratively evaluates the agent on a curated prior dataset to construct a systematic profile of its capability frontier across these two dimensions. With this profile, the probability distribution of task difficulty is adaptively computed, from which the target difficulty for the next round of training can be sampled. Guided by the sampled difficulty, a multi-agent controllable generator is finally used to synthesize high-quality interaction trajectories along with corresponding task instructions. Extensive experiments show that MobileGen consistently outperforms existing data generation methods by improving the average performance of GUI agents by 1.57 times across multiple challenging benchmarks. This highlights the importance of capability-aligned data generation for effective mobile GUI agent training.

cs.AI↗

arXiv2025

Grounding Large Language Models as Generalizable Policies in Network Control

Designing generalizable control policies that operate reliably under changing conditions is essential for robust network services in modern digital infrastructure. Yet network control remains dominated by specialized policies built from handcrafted rules or deep learning models, which struggle to generalize under real-world dynamics. Large language models (LLMs) offer a promising alternative because of their broad pretrained knowledge and emergent generalization abilities, but their practical adoption in network control is hindered by non-textual observations, constrained action spaces, complex optimization knowledge, and strict real-time requirements. Therefore, we introduce Trailblazer, a systematic framework that combines domain alignment to adapt LLMs for network control with adaptive policy collaboration to reduce inference overhead. Simulations across two heterogeneous network control tasks, adaptive bitrate streaming and cluster job scheduling, show that Trailblazer improves performance over conventional policies by 6.5%-36.6% and 3.5%-41.3%, respectively. Moreover, in a large-scale online A/B test of congestion control on Douyin, Trailblazer outperforms a highly optimized industrial policy, corresponding to a projected reduction of approximately 3,145 hours of platform-wide video stall time per day. Further analysis reveals that effective LLM-based control can be achieved by properly aligning compact LLMs and only invoking them for complex conditions, rather than model scaling or extensive invocation. Together, our results establish an LLM-driven paradigm for designing generalizable network policies and offer practical insights into grounding foundation models for real-world network control.

cs.LG↗

arXiv2025

IGen: Scalable Data Generation for Robot Learning from Open-World Images

The rise of generalist robotic policies has created an exponential demand for large-scale training data. However, on-robot data collection is labor-intensive and often limited to specific environments. In contrast, open-world images capture a vast diversity of real-world scenes that naturally align with robotic manipulation tasks, offering a promising avenue for low-cost, large-scale robot data acquisition. Despite this potential, the lack of associated robot actions hinders the practical use of open-world images for robot learning, leaving this rich visual resource largely unexploited. To bridge this gap, we propose IGen, a framework that scalably generates realistic visual observations and executable actions from open-world images. IGen first converts unstructured 2D pixels into structured 3D scene representations suitable for scene understanding and manipulation. It then leverages the reasoning capabilities of vision-language models to transform scene-specific task instructions into high-level plans and generate low-level actions as SE(3) end-effector pose sequences. From these poses, it synthesizes dynamic scene evolution and renders temporally coherent visual observations. Experiments validate the high quality of visuomotor data generated by IGen, and show that policies trained solely on IGen-synthesized data achieve performance comparable to those trained on real-world data. This highlights the potential of IGen to support scalable data generation from open-world images for generalist robotic policy training.

cs.RO↗

arXiv2025

Collaborative Belief Reasoning with LLMs for Efficient Multi-Agent Collaboration

Effective real-world multi-agent collaboration requires not only accurate planning but also the ability to reason about collaborators' intents--a crucial capability for avoiding miscoordination and redundant communication under partial observable environments. Due to their strong planning and reasoning capabilities, large language models (LLMs) have emerged as promising autonomous agents for collaborative task solving. However, existing collaboration frameworks for LLMs overlook their reasoning potential for dynamic intent inference, and thus produce inconsistent plans and redundant communication, reducing collaboration efficiency. To bridge this gap, we propose CoBel-World, a novel framework that equips LLM agents with a Collaborative Belief World--an internal representation jointly modeling the physical environment and collaborators' mental states. CoBel-World enables agents to parse external open-world knowledge into structured beliefs via a symbolic belief representation module, and perform zero-shot Bayesian-style belief updates through LLM reasoning. This allows agents to proactively detect potential miscoordination (e.g., conflicting plans) and communicate adaptively. Evaluated on challenging embodied benchmarks (i.e., TDW-MAT and C-WAH), CoBel-World significantly reduces communication costs by 64-79% and improves task completion efficiency by 4-28% compared to the strongest baseline. Our results show that explicit, intent-aware belief modeling is essential for efficient and human-like collaboration in LLM-based multi-agent systems.

cs.AI↗

arXiv2025

6G communications through sub-Terahertz CMOS power amplifiers: Design challenges and trends

The fifth-generation (5G) network faces limitations in supporting emerging applications, such as artificial intelligence (AI), virtual reality (VR) and digital twins. To overcome these confines, sub-Terahertz (sub-THz) and Terahertz (THz) technologies are considered to be key enablers of effective 6G wireless communications, offering higher transmission speeds, longer range and wider bandwidth. Achieving these capabilities requires careful engineering of 6G transceivers, with a focus on efficient power amplifiers (PAs) in the front-end, which play a critical role in effectively amplifying and transmitting signals over long distances. Complimentary metal-oxidesemiconductor (CMOS) technology-based PA in sub-THz suffers severe parasitic and limited maximum frequency, however, this has eventually been solved by different design architectures and scaling down of CMOS technology to break through the frequency limitations. In this article, we reviewed the potentials and capabilities of CMOS technology for designing 6G hardware, identified the state-of-art PA designs in the sub-THz band and then examined as well as compared the designs to identify the suitable design strategies for better performance. The circuit optimisation techniques, such as coupled-line, passive gain boosting method, zero-degree power splitting, load-pull matching, diode and capacitor linearisation for better gain, saturated output power and power added efficiency, are considered for the PA design architectures at different sub-THz bands. Furthermore, these methods are summarised and discussed with their advantages and disadvantages in lieu with their performances. The PA design trends, challenges and future perspectives are also presented and discussed. Therefore, this comprehensive review article will serve as a comparative study and reference for future PA designs for radio frequency integrated circuits (RFIC).

cs.NI↗

arXiv2024

CATP-LLM: Empowering Large Language Models for Cost-Aware Tool Planning

Utilizing large language models (LLMs) for tool planning has emerged as a promising avenue for developing general AI systems, where LLMs automatically schedule external tools (e.g., vision models) to tackle complex tasks based on task descriptions. To push this paradigm toward practical applications, it is crucial for LLMs to consider tool execution costs (e.g., execution time) for tool planning. Unfortunately, prior studies overlook the tool execution costs, leading to the generation of expensive plans whose costs outweigh their benefits in terms of task performance. To fill this gap, we propose the Cost-Aware Tool Planning with LLMs (CATP-LLM) framework, which for the first time provides a coherent design to empower LLMs for cost-aware tool planning. Specifically, To facilitate efficient concurrent tool execution and cost reduction, we design a tool planning language to enhance the LLM for creating multi-branch non-sequential plans. Moreover, we propose a cost-aware offline reinforcement learning algorithm to fine-tune the LLM to optimize the performance-cost trade-off in tool planning. In the lack of public cost-related datasets, we further present OpenCATP, the first dataset for cost-aware planning, which comprises 11,100 evaluation samples from diverse tasks. Extensive experiments show that CATP-LLM outperforms GPT-4 even when using Llama2-7B as its backbone, with the average improvement of 1.5%-93.9% in terms of plan quality. Codes and dataset are available at: https://github.com/duowuyms/OpenCATP-LLM.

cs.AI↗

arXiv2024

MoCo: Fuzzing Deep Learning Libraries via Assembling Code

The rapidly developing deep learning (DL) techniques have been applied in software systems with various application scenarios. However, they could also pose new safety threats with potentially serious consequences, especially in safety-critical domains. DL libraries serve as the underlying foundation for DL systems, and bugs in them can have unpredictable impacts that directly affect the behaviors of DL systems. Previous research on fuzzing DL libraries still has limitations in the diversity of test inputs, the construction of test oracles, and the precision of detection. In this paper, we propose MoCo, a novel fuzzing testing method for DL libraries via assembling code. MoCo first disassembles the seed code file to obtain the template and code blocks, and then employs code block mutation operators (e.g., API replacement, random generation, and boundary checking) to generate more new code blocks adapted to the template. By inserting context-appropriate code blocks into the template step by step, MoCo can generate a tree of code files with intergenerational relations. According to the derivation relations in this tree and the applied mutation operators, we construct the test oracle based on the execution state consistency. Since the granularity of code assembly and mutation is controlled rather than randomly divergent, we can quickly pinpoint the lines of code where the bugs are located and the corresponding triggering conditions. We conduct a comprehensive experiment to evaluate the efficiency and effectiveness of MoCo using three widely-used DL libraries (i.e., TensorFlow, PyTorch, and Jittor). During the experiment, MoCo detects 64 new bugs of four types in three DL libraries, where 51 bugs have been confirmed, and 13 bugs have been fixed by developers.

cs.SE↗

arXiv2024

Design of a W-band High-PAE Class A&AB Power Amplifier in 150nm GaAs Technology

Nanometer scale power amplifiers (PA) at sub-THz suffer from severe parasitic effects that lead to experience limited maximum frequency and reduced power performance at the device transceiver front end. The integrated circuits researchers proposed different PA design architecture combinations at scaled down technologies to overcome these limitations. Although the designs meet the minimum requirements, the power added efficiency (PAE) of PA is still quite low. In this paper, a W-band single-ended common-source (CS) and cascode integrated 3-stage 2-way PA design is proposed. The design integrated different key design methodologies to mitigate the parasitic; such as combined Class AB and Class A stages for gain-boosting and efficiency enhancement, Wilkinson power combiner for higher output power, linearity, and bandwidth, and transmission line (TL)-based wide band matching network for better inter-stage matching and compact size. The proposed PA design is validated using UMS 150-nm GaAs pHEMT using advanced design system (ADS) simulator. The results show that the proposed PA achieved a gain of 20.1 dB, an output power of 17.2 dBm, a PAE of 33 % and a 21 GHz bandwidth at 90 GHz Sub-THz band. The PA layout consumes only 5.66 X 2.51 mm2 die space including pads. Our proposed PA design will boost the research on sub-THz integrated circuits research and will smooth the wide spread adoption of 6G in near future.

cs.NI↗

arXiv2024

NetLLM: Adapting Large Language Models for Networking

Many networking tasks now employ deep learning (DL) to solve complex prediction and optimization problems. However, current design philosophy of DL-based algorithms entails intensive engineering overhead due to the manual design of deep neural networks (DNNs) for different networking tasks. Besides, DNNs tend to achieve poor generalization performance on unseen data distributions/environments. Motivated by the recent success of large language models (LLMs), this work studies the LLM adaptation for networking to explore a more sustainable design philosophy. With the powerful pre-trained knowledge, the LLM is promising to serve as the foundation model to achieve "one model for all tasks" with even better performance and stronger generalization. In pursuit of this vision, we present NetLLM, the first framework that provides a coherent design to harness the powerful capabilities of LLMs with low efforts to solve networking problems. Specifically, NetLLM empowers the LLM to effectively process multimodal data in networking and efficiently generate task-specific answers. Besides, NetLLM drastically reduces the costs of fine-tuning the LLM to acquire domain knowledge for networking. Across three networking-related use cases - viewport prediction, adaptive bitrate streaming and cluster job scheduling, we showcase that the NetLLM-adapted LLM significantly outperforms state-of-the-art algorithms.

cs.NI↗

arXiv2023

MANSY: Generalizing Neural Adaptive Immersive Video Streaming With Ensemble and Representation Learning

The popularity of immersive videos has prompted extensive research into neural adaptive tile-based streaming to optimize video transmission over networks with limited bandwidth. However, the diversity of users' viewing patterns and Quality of Experience (QoE) preferences has not been fully addressed yet by existing neural adaptive approaches for viewport prediction and bitrate selection. Their performance can significantly deteriorate when users' actual viewing patterns and QoE preferences differ considerably from those observed during the training phase, resulting in poor generalization. In this paper, we propose MANSY, a novel streaming system that embraces user diversity to improve generalization. Specifically, to accommodate users' diverse viewing patterns, we design a Transformer-based viewport prediction model with an efficient multi-viewport trajectory input output architecture based on implicit ensemble learning. Besides, we for the first time combine the advanced representation learning and deep reinforcement learning to train the bitrate selection model to maximize diverse QoE objectives, enabling the model to generalize across users with diverse preferences. Extensive experiments demonstrate that MANSY outperforms state-of-the-art approaches in viewport prediction accuracy and QoE improvement on both trained and unseen viewing patterns and QoE preferences, achieving better generalization.

cs.NI↗

arXiv2023

ILCAS: Imitation Learning-Based Configuration-Adaptive Streaming for Live Video Analytics with Cross-Camera Collaboration

The high-accuracy and resource-intensive deep neural networks (DNNs) have been widely adopted by live video analytics (VA), where camera videos are streamed over the network to resource-rich edge/cloud servers for DNN inference. Common video encoding configurations (e.g., resolution and frame rate) have been identified with significant impacts on striking the balance between bandwidth consumption and inference accuracy and therefore their adaption scheme has been a focus of optimization. However, previous profiling-based solutions suffer from high profiling cost, while existing deep reinforcement learning (DRL) based solutions may achieve poor performance due to the usage of fixed reward function for training the agent, which fails to craft the application goals in various scenarios. In this paper, we propose ILCAS, the first imitation learning (IL) based configuration-adaptive VA streaming system. Unlike DRL-based solutions, ILCAS trains the agent with demonstrations collected from the expert which is designed as an offline optimal policy that solves the configuration adaption problem through dynamic programming. To tackle the challenge of video content dynamics, ILCAS derives motion feature maps based on motion vectors which allow ILCAS to visually ``perceive'' video content changes. Moreover, ILCAS incorporates a cross-camera collaboration scheme to exploit the spatio-temporal correlations of cameras for more proper configuration selection. Extensive experiments confirm the superiority of ILCAS compared with state-of-the-art solutions, with 2-20.9% improvement of mean accuracy and 19.9-85.3% reduction of chunk upload lag.

cs.NI↗