arXiv subjects

Wenfeng Li

Publications and source records attributed to Wenfeng Li.

12 recordsLinked to original sources

arXiv2026

Bioinfoysis Technical Report

Large language model agents have shown promise in bioinformatics, but most existing systems focus primarily on producing final answers, treating planning, tool use, and code execution as transient interactions. This design is poorly suited to long-horizon bioinformatics tasks, where conclusions must remain connected to the data, computations, and intermediate evidence that support them. We introduce \textbf{Bioinfoysis}, a multi-agent harness that represents each request as a persistent, artifact-grounded analysis run. Bioinfoysis combines global planning with step-wise, evidence-driven replanning: the planner maintains an executable checklist and revises pending steps using structured handoffs returned after each worker execution. These handoffs bind intermediate results to their responsible agent, checklist step, and plan generation, preventing stale evidence from being silently reused after replanning. A controlled runtime validates generated scripts, tables, and figures before they are used in downstream analysis or reporting, while role-specific context, persistent memory, and governed bioinformatics skills support reliable execution over long analysis trajectories. We evaluate Bioinfoysis on BixBench and two question-answering tracks of LAB-Bench 2. On BixBench, Bioinfoysis achieves state-of-the-art accuracy of 82.4\%. Across four underlying language models, Bioinfoysis increases average accuracy from 27.81\% to 64.13\% on SeqQA2 and from 3.13\% to 31.25\% on DbQA2. These results demonstrate that reliable bioinformatics automation depends not only on model capability, but also on the harness that governs planning, execution, memory, and evidence flow. We hope that the emergence of Bioinfoysis will play a driving and leading role in the development of the bioinformatics community. Our demo website can be seen in https://report.bioinfoysis.com/.

cs.AI↗

arXiv2026

SafeFlow: Semantic Information-Flow Control for Blocking Malicious Propagation in Multi-Agent Systems

Multi-agent systems improve capability through task decomposition and role specialization, but these same mechanisms introduce an important safety blind spot: a harmful objective can be fragmented into locally plausible subtasks, allowing malicious intent to evade detection by any single agent. This is a growing social-impact challenge: systems handling sensitive information or consequential tools can turn routine delegation into unauthorized disclosure or unsafe action. We argue that this failure mode is better understood as a semantic information-flow problem than as a single-turn prompt classification task. To address this, we propose SafeFlow, a defense framework for multi-agent systems that formalizes malicious cross-agent propagation as a semantic information-flow problem. SafeFlow attaches structured semantic taints to root requests, propagates them through a dynamic collaboration graph, and performs workflow-level validation to reconstruct the global risk context before irreversible actions are committed. Evaluated on four benchmarks spanning prompt injection, jailbreak-based unsafe tool use, risky code execution, and harmful web-agent behavior, SafeFlow reduces attack success rates compared to undefended baselines and external defenses while retaining high benign task completion and a high paired safe--harm success rate. Our findings show that multi-agent systems still lack mechanisms for preserving risk semantics across delegation boundaries. This gap can turn routine delegation into privacy harms or unsafe actions that affect people and organizations. SafeFlow keeps this risk visible throughout the workflow, before it results in harm.

cs.MA↗

arXiv2026

Progressive Replacement Sequence Planning for Laser-Enhanced BeiDou Navigation Constellations

The early satellites of the third-generation BeiDou Navigation Satellite System (BDS-3) are approaching the end of their design lifetime, making progressive constellation replacement an inevitable engineering task. Meanwhile, laser inter-satellite links (LISLs) provide high-precision time transfer, sub-millimeter-level ranging, and high-rate data forwarding capabilities, offering a promising upgrade path for future BeiDou satellites. This paper investigates the replacement-sequence planning problem for progressively replacing legacy microwave satellites with laser-enabled satellites from a networking perspective. To address this problem, three main contributions are made: (i) BeiDou progressive replacement is formulated as a network-aware launch sequence planning problem over evolving laser--microwave hybrid constellation states; (ii) an integer linear programming (ILP) model is developed to evaluate the networking gain of each candidate launch action under practical engineering constraints; and (iii) a priority-aware search heuristic is introduced to reduce the action evaluation space for efficient round-wise decision making. Simulations based on the actual BDS-3 constellation demonstrate that the proposed replacement sequence achieves a more favorable hybrid-network evolution than the historical BDS-3 launch order, including earlier formation of a ground-accessible laser-enhanced network structure, better utilization of laser-link resources, fewer residual non-anchor satellites, and lower microwave-layer waiting delay. The proposed framework can serve as a network-aware decision-support tool for future BeiDou evolution, and the resulting replacement sequence can provide direct engineering guidance for BeiDou replacement planning.

eess.SY↗

arXiv2026

A Secure Splitting and Acceleration Strategy for TCP/QUIC in Interplanetary Networks

Interplanetary networks (IPNs) present unique challenges such as extreme delay, high loss, and frequent disruptions that severely degrade the performance of conventional transport protocols like Transmission Control Protocol (TCP) and Quick UDP Internet Connection (QUIC). To address these issues, we propose a secure transport acceleration strategy tailored for IPNs. This strategy is founded on our Non-Transparent Secure Proxy (NTSP) architecture, which enables connection splitting for end-to-end encrypted flows while preserving application layer security. Based on the NTSP, we design an IPN-aware transport policy that combines (i) a rate-based congestion control algorithm exploiting the pre-scheduled nature of deep-space links to achieve stable and efficient bandwidth utilization, and (ii) an adaptive packet-level forward error correction scheme to provide low-latency loss recovery without retransmissions. Furthermore, we introduce a theoretically grounded backpressure flow control mechanism, deriving an analytical model for optimal buffer sizing to mitigate rate mismatch and prevent bufferbloat. The strategy is implemented in a prototype system, PEPspace, and evaluated in representative Earth-Moon scenarios. Results show near-capacity and stable goodput and substantially improved delivery performance compared with TCP/QUIC variants and existing Performance Enhancing Proxies, while maintaining low latency and robust data delivery across intermittent links. The NTSP architecture is further discussed as a foundational framework for future unified IP/DTN architectures, bridging a key architectural gap in heterogeneous space networks.

cs.NI↗

arXiv2026

Contact Plan Design for Cross-Linked GNSSs: An ILP Approach for Extended Applications

Global Navigation Satellite Systems (GNSS) employ inter-satellite links (ISLs) to reduce dependency on ground stations, enabling precise ranging and communication across satellites. Beyond their traditional role, ISLs can support extended applications, including providing navigation and communication services to external entities. However, designing effective contact plan design (CPD) schemes for these multifaceted ISLs, operating under a polling time-division duplex (PTDD) framework, remains a critical challenge. Existing CPD approaches focus solely on meeting GNSS satellites' internal ranging and communication demands, neglecting their extended applications. This paper introduces the first CPD scheme capable of supporting extended GNSS ISLs. By modeling GNSS requirements and designing a tailored service process, our approach ensures the allocation of essential resources for internal operations while accommodating external user demands. Based on the BeiDou constellation, simulation results demonstrate the proposed scheme's efficacy in maintaining core GNSS functionality while providing extended ISLs on a best-effort basis. Additionally, the results highlight the significant impact of GNSS ISLs in enhancing orbit determination and clock synchronization for the Earth-Moon libration point constellation, underscoring the importance of extended GNSS ISL applications.

eess.SY↗

arXiv2025

Optimized Contact Plan Design for Reflector and Phased Array Terminals in Cislunar Space Networks

Cislunar space is emerging as a critical domain for human exploration, requiring robust infrastructure to support spatial users-spacecraft with navigation and communication demands. Deploying satellites at Earth-Moon three-body orbits offers an effective solution to construct cislunar space infrastructure (CLSI). However, scheduling satellite links to serve users necessitates an appropriate contact plan design (CPD) scheme. Existing CPD schemes focus solely on inter-satellite link scheduling, overlooking their role in providing services to users. This paper introduces a CPD scheme that considers two classes of satellite transponders: Reflector Links (RL) for high-volume data transfers and Phased Array Links (PL) for fast switching and navigation services. Our approach supports both satellites and spatial users in cislunar space. Simulations validate the scheme, demonstrating effective support for user while meeting satellite ranging and communication requirements. These findings provide essential insights for developing future Cislunar Space Infrastructures.

eess.SY↗

arXiv2025

Joint Contact Planning for Navigation and Communication in GNSS-Libration Point Systems

Deploying satellites at Earth-Moon Libration Points (LPs) addresses the inherent deep-space coverage gaps of low-altitude GNSS constellations. Integrating LP satellites with GNSS into a joint constellation enables a more robust and comprehensive Positioning, Navigation, and Timing (PNT) system, while also extending navigation and communication services to spacecraft operating in cislunar space (i.e., users). However, the long propagation delays between LP satellites, users, and GNSS satellites result in significantly different link durations compared to those within the GNSS constellation. Scheduling inter-satellite links (ISLs) is a core task of Contact Plan Design (CPD). Existing CPD approaches focus exclusively on GNSS constellations, assuming uniform link durations, and thus cannot accommodate the heterogeneous link timescales present in a joint GNSS-LP system. To overcome this limitation, we introduce a Joint CPD (J-CPD) scheme tailored to handle ISLs with differing duration units across integrated constellations. The key contributions of J-CPD are: (i):introduction of LongSlots (Earth-Moon scale links) and ShortSlots (GNSS-scale links); (ii):a hierarchical and crossed CPD process for scheduling LongSlots and ShortSlots ISLs; (iii):an energy-driven link scheduling algorithm adapted to the CPD process. Simulations on a joint BeiDou-LP constellation demonstrate that J-CPD surpasses the baseline FCP method in both delay and ranging coverage, while maintaining high user satisfaction and enabling tunable trade-offs through adjustable potential-energy parameters. To our knowledge, this is the first CPD framework to jointly optimize navigation and communication in GNSS-LP systems, representing a key step toward unified and resilient deep-space PNT architectures.

eess.SY↗

arXiv2025

Reliable Transmission of LTP Using Reinforcement Learning-Based Adaptive FEC

Delay/Disruption Tolerant Networking (DTN) employs the Licklider Transmission Protocol (LTP) with Automatic Repeat reQuest (ARQ) for reliable data delivery in challenging interplanetary networks. While previous studies have integrated packet-level Forward Erasure Correction (FEC) into LTP to reduce retransmission time costs, existing static and delay-feedback-based dynamic coding methods struggle with highly variable and unpredictable deep space channel conditions. This paper proposes a reinforcement learning (RL)-based adaptive FEC algorithm to address these limitations. The algorithm utilizes historical feedback and system state to predict future channel conditions and proactively adjust the code rate. This approach aims to anticipate channel quality degradation, thereby preventing decoding failures and subsequent LTP retransmissions and improving coding efficiency by minimizing redundancy during favorable channel conditions. Performance evaluations conducted in simulated Earth-Moon and Earth-Mars link scenarios demonstrate this algorithm's effectiveness in optimizing data transmission for interplanetary networks. Compared to existing methods, this approach demonstrates significant improvement, with matrix decoding failures reduced by at least 2/3.

cs.NI↗

arXiv2024

Cooperative Classification and Rationalization for Graph Generalization

Graph Neural Networks (GNNs) have achieved impressive results in graph classification tasks, but they struggle to generalize effectively when faced with out-of-distribution (OOD) data. Several approaches have been proposed to address this problem. Among them, one solution is to diversify training distributions in vanilla classification by modifying the data environment, yet accessing the environment information is complex. Besides, another promising approach involves rationalization, extracting invariant rationales for predictions. However, extracting rationales is difficult due to limited learning signals, resulting in less accurate rationales and diminished predictions. To address these challenges, in this paper, we propose a Cooperative Classification and Rationalization (C2R) method, consisting of the classification and the rationalization module. Specifically, we first assume that multiple environments are available in the classification module. Then, we introduce diverse training distributions using an environment-conditional generative network, enabling robust graph representations. Meanwhile, the rationalization module employs a separator to identify relevant rationale subgraphs while the remaining non-rationale subgraphs are de-correlated with labels. Next, we align graph representations from the classification module with rationale subgraph representations using the knowledge distillation methods, enhancing the learning signal for rationales. Finally, we infer multiple environments by gathering non-rationale representations and incorporate them into the classification module for cooperative learning. Extensive experimental results on both benchmarks and synthetic datasets demonstrate the effectiveness of C2R. Code is available at https://github.com/yuelinan/Codes-of-C2R.

cs.LG↗

arXiv2024

Text2MDT: Extracting Medical Decision Trees from Medical Texts

Knowledge of the medical decision process, which can be modeled as medical decision trees (MDTs), is critical to build clinical decision support systems. However, the current MDT construction methods rely heavily on time-consuming and laborious manual annotation. In this work, we propose a novel task, Text2MDT, to explore the automatic extraction of MDTs from medical texts such as medical guidelines and textbooks. We normalize the form of the MDT and create an annotated Text-to-MDT dataset in Chinese with the participation of medical experts. We investigate two different methods for the Text2MDT tasks: (a) an end-to-end framework which only relies on a GPT style large language models (LLM) instruction tuning to generate all the node information and tree structures. (b) The pipeline framework which decomposes the Text2MDT task to three subtasks. Experiments on our Text2MDT dataset demonstrate that: (a) the end-to-end method basd on LLMs (7B parameters or larger) show promising results, and successfully outperform the pipeline methods. (b) The chain-of-thought (COT) prompting method \cite{Wei2022ChainOT} can improve the performance of the fine-tuned LLMs on the Text2MDT test set. (c) the lightweight pipelined method based on encoder-based pretrained models can perform comparably with LLMs with model complexity two magnititudes smaller. Our Text2MDT dataset is open-sourced at \url{https://tianchi.aliyun.com/dataset/95414}, and the source codes are open-sourced at \url{https://github.com/michael-wzhu/text2dt}.

cs.CL↗

arXiv2021

AutoKWS: Keyword Spotting with Differentiable Architecture Search

Smart audio devices are gated by an always-on lightweight keyword spotting program to reduce power consumption. It is however challenging to design models that have both high accuracy and low latency for accurate and fast responsiveness. Many efforts have been made to develop end-to-end neural networks, in which depthwise separable convolutions, temporal convolutions, and LSTMs are adopted as building units. Nonetheless, these networks designed with human expertise may not achieve an optimal trade-off in an expansive search space. In this paper, we propose to leverage recent advances in differentiable neural architecture search to discover more efficient networks. Our searched model attains 97.2% top-1 accuracy on Google Speech Command Dataset v1 with only nearly 100K parameters.

eess.AS↗

arXiv2012

Sublinear Time Approximate Sum via Uniform Random Sampling

We investigate the approximation for computing the sum $a_1+...+a_n$ with an input of a list of nonnegative elements $a_1,..., a_n$. If all elements are in the range $[0,1]$, there is a randomized algorithm that can compute an $(1+ε)$-approximation for the sum problem in time ${O({n(\log\log n)\over\sum_{i=1}^n a_i})}$, where $ε$ is a constant in $(0,1)$. Our randomized algorithm is based on the uniform random sampling, which selects one element with equal probability from the input list each time. We also prove a lower bound $Ω({n\over \sum_{i=1}^n a_i})$, which almost matches the upper bound, for this problem.

cs.DS↗