Yepeng Ding

Ahmed Twabi, Yepeng Ding, Tohru Kondo

As agentic network management gains popularity, there is a critical need for evaluation frameworks that transcend static, one-shot testing. To address this, we introduce NetAgentBench, a dynamic benchmark that evaluates agent interactions through a Finite State Machine (FSM) formalization guaranteeing determinism, correctness, and bounded execution. This provides the networking landscape with a rigorous foundation to measure complex, multi-turn operational behaviors. Our empirical evaluation of four state-of-the-art LLM agents through diverse network configuration tasks reveals stark deficiencies: while agents can solve basic tasks, they suffer severe exploration meltdowns and coherence collapse during expert-level configurations. Ultimately, NetAgentBench demonstrates that systematically evaluating multi-turn behavioral stability is an indispensable step toward realizing trustworthy, fully autonomous networks.

Junwei Yu, Mufeng Yang, Yepeng Ding et al.

The proliferation of AI-powered search engines has shifted information discovery from traditional link-based retrieval to direct answer generation with selective source citation, creating new challenges for content visibility. While existing Generative Engine Optimization (GEO) approaches focus primarily on semantic content modification, the role of structural features in influencing citation behavior remains underexplored. In this paper, we propose GEO-SFE, a systematic framework for structural feature engineering in generative engine optimization. Our approach decomposes content structure into three hierarchical levels: macro-structure (document architecture), meso-structure (information chunking), and micro-structure (visual emphasis), and models their impact on citation probability across different generative engine architectures. We develop architecture-aware optimization strategies and predictive models that preserve semantic integrity while improving structural effectiveness. Experimental evaluation across six mainstream generative engines demonstrates consistent improvements in citation rate (17.3 percent) and subjective quality (18.5 percent), validating the effectiveness and generalizability of the proposed framework. This work establishes structural optimization as a foundational component of GEO, providing a data-driven methodology for enhancing content visibility in LLM-powered information ecosystems.

Junwei Yu, Yepeng Ding, Hiroyuki Sato

The emergence of Large Language Models (LLMs) in Multi-Agent Systems (MAS) has opened new possibilities for artificial intelligence, yet current implementations face significant challenges in resource management, task coordination, and system efficiency. While existing frameworks demonstrate the potential of LLM-based agents in collaborative problem-solving, they often lack sophisticated mechanisms for parallel execution and dynamic task management. This paper introduces DynTaskMAS, a novel framework that orchestrates asynchronous and parallel operations in LLM-based MAS through dynamic task graphs. The framework features four key innovations: (1) a Dynamic Task Graph Generator that intelligently decomposes complex tasks while maintaining logical dependencies, (2) an Asynchronous Parallel Execution Engine that optimizes resource utilization through efficient task scheduling, (3) a Semantic-Aware Context Management System that enables efficient information sharing among agents, and (4) an Adaptive Workflow Manager that dynamically optimizes system performance. Experimental evaluations demonstrate that DynTaskMAS achieves significant improvements over traditional approaches: a 21-33% reduction in execution time across task complexities (with higher gains for more complex tasks), a 35.4% improvement in resource utilization (from 65% to 88%), and near-linear throughput scaling up to 16 concurrent agents (3.47X improvement for 4X agents). Our framework establishes a foundation for building scalable, high-performance LLM-based multi-agent systems capable of handling complex, dynamic tasks efficiently.

Lingfeng Zhang, Yuheng Guo, Yepeng Ding et al.

Online signature verification plays a pivotal role in security infrastructures. However, conventional online signature verification models pose significant risks to data privacy, especially during training processes. To mitigate these concerns, we propose a novel federated learning framework that leverages 1-D Convolutional Neural Networks (CNN) for online signature verification. Furthermore, our experiments demonstrate the effectiveness of our framework regarding 1-D CNN and federated learning. Particularly, the experiment results highlight that our framework 1) minimizes local computational resources; 2) enhances transfer effects with substantial initialization data; 3) presents remarkable scalability. The centralized 1-D CNN model achieves an Equal Error Rate (EER) of 3.33% and an accuracy of 96.25%. Meanwhile, configurations with 2, 5, and 10 agents yield EERs of 5.42%, 5.83%, and 5.63%, along with accuracies of 95.21%, 94.17%, and 94.06%, respectively.

Yepeng Ding, Hiroyuki Sato

Blockchain technologies have been boosting the development of data-driven decentralized services in a wide range of fields. However, with the spirit of full transparency, many public blockchains expose all types of data to the public such as Ethereum. Besides, the on-chain persistence of large data is significantly expensive technically and economically. These issues lead to the difficulty of sharing fairly large private data while preserving attractive properties of public blockchains. Although direct encryption for on-chain data persistence can introduce confidentiality, new challenges such as key sharing, access control, and legal rights proving are still open. Meanwhile, cross-chain collaboration still requires secure and effective protocols, though decentralized storage systems such as IPFS bring the possibility for fairly large data persistence. In this paper, we propose Sunspot, a decentralized framework for privacy-preserving data sharing with access control on transparent public blockchains, to solve these issues. We also show the practicality and applicability of Sunspot by MyPub, a decentralized privacy-preserving publishing platform based on Sunspot. Furthermore, we evaluate the security, privacy, and performance of Sunspot through theoretical analysis and experiments.

Yepeng Ding, Hiroyuki Sato

Decentralized systems have been widely developed and applied to address security and privacy issues in centralized systems, especially since the advancement of distributed ledger technology. However, it is challenging to ensure their correct functioning with respect to their designs and minimize the technical risk before the delivery. Although formal methods have made significant progress over the past decades, a feasible solution based on formal methods from a development process perspective has not been well developed. In this paper, we formulate an iterative and incremental development process, named formalism-driven development (FDD), for developing provably correct decentralized systems under the guidance of formal methods. We also present a framework named Seniz, to practicalize FDD with a new modeling language and scaffolds. Furthermore, we conduct case studies to demonstrate the effectiveness of FDD in practice with the support of Seniz.