arXiv subjects

Wenjing Zhang

Publications and source records attributed to Wenjing Zhang.

At least 55 records · Page 3Linked to original sources

arXiv2025

ToMacVF : Temporal Macro-action Value Factorization for Asynchronous Multi-Agent Reinforcement Learning

Existing asynchronous MARL methods based on MacDec-POMDP typically construct training trajectory buffers by simply sampling limited and biased data at the endpoints of macro-actions, and directly apply conventional MARL methods on the buffers. As a result, these methods lead to an incomplete and inaccurate representation of the macro-action execution process, along with unsuitable credit assignments. To solve these problems, the Temporal Macro-action Value Factorization (ToMacVF) is proposed to achieve fine-grained temporal credit assignment for macro-action contributions. A centralized training buffer, called Macro-action Segmented Joint Experience Replay Trajectory (Mac-SJERT), is designed to incorporate with ToMacVF to collect accurate and complete macro-action execution information, supporting a more comprehensive and precise representation of the macro-action process. To ensure principled and fine-grained asynchronous value factorization, the consistency requirement between joint and individual macro-action selection called Temporal Macro-action based IGM (To-Mac-IGM) is formalized, proving that it generalizes the synchronous cases. Based on To-Mac-IGM, a modularized ToMacVF architecture, which satisfies CTDE principle, is designed to conveniently integrate previous value factorization methods. Next, the ToMacVF algorithm is devised as an implementation of the ToMacVF architecture. Experimental results demonstrate that, compared to asynchronous baselines, our ToMacVF algorithm not only achieves optimal performance but also exhibits strong adaptability and robustness across various asynchronous multi-agent experimental scenarios.

cs.MA↗

arXiv2025

A Scalable and Efficient Signal Integration System for Job Matching

LinkedIn, one of the world's largest platforms for professional networking and job seeking, encounters various modeling challenges in building recommendation systems for its job matching product, including cold-start, filter bubbles, and biases affecting candidate-job matching. To address these, we developed the STAR (Signal Integration for Talent And Recruiters) system, leveraging the combined strengths of Large Language Models (LLMs) and Graph Neural Networks (GNNs). LLMs excel at understanding textual data, such as member profiles and job postings, while GNNs capture intricate relationships and mitigate cold-start issues through network effects. STAR integrates diverse signals by uniting LLM and GNN capabilities with industrial-scale paradigms including adaptive sampling and version management. It provides an end-to-end solution for developing and deploying embeddings in large-scale recommender systems. Our key contributions include a robust methodology for building embeddings in industrial applications, a scalable GNN-LLM integration for high-performing recommendations, and practical insights for real-world model deployment.

cs.LG↗

arXiv2025

Zero-Shot Semantic Communication with Multimodal Foundation Models

Most existing semantic communication (SemCom) systems use deep joint source-channel coding (DeepJSCC) to encode task-specific semantics in a goal-oriented manner. However, their reliance on predefined tasks and datasets significantly limits their flexibility and generalizability in practical deployments. Multi-modal foundation models provide a promising solution by generating universal semantic tokens. Inspired by this, we introduce SemCLIP, a zero-shot SemCom framework leveraging the contrastive language-image pre-training (CLIP) model. By transmitting CLIP-generated image tokens instead of raw images, SemCLIP enables efficient SemCom under low bandwidth and challenging channel conditions, facilitating diverse downstream tasks and zero-shot applications. Specifically, we propose a DeepJSCC scheme for efficient CLIP token encoding. To mitigate potential degradation caused by compression and channel noise, a multi-modal transmission-aware prompt learning mechanism is designed at the receiver, which adapts prompts based on transmission quality, enhancing system robustness and channel adaptability. Simulation results demonstrate that SemCLIP outperforms the baselines, achieving a $41\%$ improvement in zero-shot performance at low signal-to-noise ratios. Meanwhile, SemCLIP reduces bandwidth usage by more than $50$-fold compared to alternative image transmission methods, demonstrating the potential of foundation models towards a generalized, task-agnostic SemCom solution.

eess.SP↗

arXiv2025

Limiting absorption principle of Helmholtz equation with sign changing coefficients under periodic structure

Negative refractive index materials have attracted significant research attention due to their unique electromagnetic response characteristics. In this paper, we employ the complementing boundary condition to establish rigorous a priori estimates for the Helmholtz equation, from which the limiting absorption principle is analytically derived. Within this mathematical framework, we conclusively establish the well-posedness of the electromagnetic transmission problem at the interface between conventional materials and negative refractive index materials in two-dimensional periodic structures.

math.AP↗

arXiv2025

Safety Evaluation and Enhancement of DeepSeek Models in Chinese Contexts

DeepSeek-R1, renowned for its exceptional reasoning capabilities and open-source strategy, is significantly influencing the global artificial intelligence landscape. However, it exhibits notable safety shortcomings. Recent research conducted by Robust Intelligence, a subsidiary of Cisco, in collaboration with the University of Pennsylvania, revealed that DeepSeek-R1 achieves a 100\% attack success rate when processing harmful prompts. Furthermore, multiple security firms and research institutions have identified critical security vulnerabilities within the model. Although China Unicom has uncovered safety vulnerabilities of R1 in Chinese contexts, the safety capabilities of the remaining distilled models in the R1 series have not yet been comprehensively evaluated. To address this gap, this study utilizes the comprehensive Chinese safety benchmark CHiSafetyBench to conduct an in-depth safety evaluation of the DeepSeek-R1 series distilled models. The objective is to assess the safety capabilities of these models in Chinese contexts both before and after distillation, and to further elucidate the adverse effects of distillation on model safety. Building on these findings, we implement targeted safety enhancements for the entire DeepSeek-R1 model series. Evaluation results indicate that the enhanced models achieve significant improvements in safety while maintaining reasoning capabilities without notable degradation. We open-source the safety-enhanced models at https://github.com/UnicomAI/DeepSeek-R1-Safe to serve as a valuable resource for future research and optimization of DeepSeek models.

cs.CL↗

arXiv2025

Safety Evaluation of DeepSeek Models in Chinese Contexts

Recently, the DeepSeek series of models, leveraging their exceptional reasoning capabilities and open-source strategy, is reshaping the global AI landscape. Despite these advantages, they exhibit significant safety deficiencies. Research conducted by Robust Intelligence, a subsidiary of Cisco, in collaboration with the University of Pennsylvania, revealed that DeepSeek-R1 has a 100\% attack success rate when processing harmful prompts. Additionally, multiple safety companies and research institutions have confirmed critical safety vulnerabilities in this model. As models demonstrating robust performance in Chinese and English, DeepSeek models require equally crucial safety assessments in both language contexts. However, current research has predominantly focused on safety evaluations in English environments, leaving a gap in comprehensive assessments of their safety performance in Chinese contexts. In response to this gap, this study introduces CHiSafetyBench, a Chinese-specific safety evaluation benchmark. This benchmark systematically evaluates the safety of DeepSeek-R1 and DeepSeek-V3 in Chinese contexts, revealing their performance across safety categories. The experimental results quantify the deficiencies of these two models in Chinese contexts, providing key insights for subsequent improvements. It should be noted that, despite our efforts to establish a comprehensive, objective, and authoritative evaluation benchmark, the selection of test samples, characteristics of data distribution, and the setting of evaluation criteria may inevitably introduce certain biases into the evaluation results. We will continuously optimize the evaluation benchmark and periodically update this report to provide more comprehensive and accurate assessment outcomes. Please refer to the latest version of the paper for the most recent evaluation results and conclusions.

cs.CL↗

arXiv2025

Universal differential equations for optimal control problems and its application on cancer therapy

This paper highlights a parallel between the forward backward sweeping method for optimal control and deep learning training procedures. We reformulate a classical optimal control problem, constrained by a differential equation system, into an optimization framework that uses neural networks to represent control variables. We demonstrate that this deep learning method adheres to Pontryagin Maximum Principle and mitigates numerical instabilities by employing backward propagation instead of a backward sweep for the adjoint equations. As a case study, we solve an optimal control problem to find the optimal combination of immunotherapy and chemotherapy. Our approach holds significant potential across various fields, including epidemiology, ecological modeling, engineering, and financial mathematics, where optimal control under complex dynamic constraints is crucial.

math.OC↗

arXiv2024

CHiSafetyBench: A Chinese Hierarchical Safety Benchmark for Large Language Models

With the profound development of large language models(LLMs), their safety concerns have garnered increasing attention. However, there is a scarcity of Chinese safety benchmarks for LLMs, and the existing safety taxonomies are inadequate, lacking comprehensive safety detection capabilities in authentic Chinese scenarios. In this work, we introduce CHiSafetyBench, a dedicated safety benchmark for evaluating LLMs' capabilities in identifying risky content and refusing answering risky questions in Chinese contexts. CHiSafetyBench incorporates a dataset that covers a hierarchical Chinese safety taxonomy consisting of 5 risk areas and 31 categories. This dataset comprises two types of tasks: multiple-choice questions and question-answering, evaluating LLMs from the perspectives of risk content identification and the ability to refuse answering risky questions respectively. Utilizing this benchmark, we validate the feasibility of automatic evaluation as a substitute for human evaluation and conduct comprehensive automatic safety assessments on mainstream Chinese LLMs. Our experiments reveal the varying performance of different models across various safety domains, indicating that all models possess considerable potential for improvement in Chinese safety capabilities. Our dataset is publicly available at https://github.com/UnicomAI/UnicomBenchmark/tree/main/CHiSafetyBench.

cs.CL↗

arXiv2024

B2MAPO: A Batch-by-Batch Multi-Agent Policy Optimization to Balance Performance and Efficiency

Most multi-agent reinforcement learning approaches adopt two types of policy optimization methods that either update policy simultaneously or sequentially. Simultaneously updating policies of all agents introduces non-stationarity problem. Although sequentially updating policies agent-by-agent in an appropriate order improves policy performance, it is prone to low efficiency due to sequential execution, resulting in longer model training and execution time. Intuitively, partitioning policies of all agents according to their interdependence and updating joint policy batch-by-batch can effectively balance performance and efficiency. However, how to determine the optimal batch partition of policies and batch updating order are challenging problems. Firstly, a sequential batched policy updating scheme, B2MAPO (Batch by Batch Multi-Agent Policy Optimization), is proposed with a theoretical guarantee of the monotonic incrementally tightened bound. Secondly, a universal modulized plug-and-play B2MAPO hierarchical framework, which satisfies CTDE principle, is designed to conveniently integrate any MARL models to fully exploit and merge their merits, including policy optimality and inference efficiency. Next, a DAG-based B2MAPO algorithm is devised, which is a carefully designed implementation of B2MAPO framework. Comprehensive experimental results conducted on StarCraftII Multi-agent Challenge and Google Football Research demonstrate the performance of DAG-based B2MAPO algorithm outperforms baseline methods. Meanwhile, compared with A2PO, our algorithm reduces the model training and execution time by 60.4% and 78.7%, respectively.

cs.MA↗

arXiv2024

Methodology of Adapting Large English Language Models for Specific Cultural Contexts

The rapid growth of large language models(LLMs) has emerged as a prominent trend in the field of artificial intelligence. However, current state-of-the-art LLMs are predominantly based on English. They encounter limitations when directly applied to tasks in specific cultural domains, due to deficiencies in domain-specific knowledge and misunderstandings caused by differences in cultural values. To address this challenge, our paper proposes a rapid adaptation method for large models in specific cultural contexts, which leverages instruction-tuning based on specific cultural knowledge and safety values data. Taking Chinese as the specific cultural context and utilizing the LLaMA3-8B as the experimental English LLM, the evaluation results demonstrate that the adapted LLM significantly enhances its capabilities in domain-specific knowledge and adaptability to safety values, while maintaining its original expertise advantages.

cs.CL↗

arXiv2024

What is the best model? Application-driven Evaluation for Large Language Models

General large language models enhanced with supervised fine-tuning and reinforcement learning from human feedback are increasingly popular in academia and industry as they generalize foundation models to various practical tasks in a prompt manner. To assist users in selecting the best model in practical application scenarios, i.e., choosing the model that meets the application requirements while minimizing cost, we introduce A-Eval, an application-driven LLMs evaluation benchmark for general large language models. First, we categorize evaluation tasks into five main categories and 27 sub-categories from a practical application perspective. Next, we construct a dataset comprising 678 question-and-answer pairs through a process of collecting, annotating, and reviewing. Then, we design an objective and effective evaluation method and evaluate a series of LLMs of different scales on A-Eval. Finally, we reveal interesting laws regarding model scale and task difficulty level and propose a feasible method for selecting the best model. Through A-Eval, we provide clear empirical and engineer guidance for selecting the best model, reducing barriers to selecting and using LLMs and promoting their application and development. Our benchmark is publicly available at https://github.com/UnicomAI/DataSet/tree/main/TestData/GeneralAbility.

cs.CL↗

arXiv2024

SMAUG: A Sliding Multidimensional Task Window-Based MARL Framework for Adaptive Real-Time Subtask Recognition

Instead of making behavioral decisions directly from the exponentially expanding joint observational-action space, subtask-based multi-agent reinforcement learning (MARL) methods enable agents to learn how to tackle different subtasks. Most existing subtask-based MARL methods are based on hierarchical reinforcement learning (HRL). However, these approaches often limit the number of subtasks, perform subtask recognition periodically, and can only identify and execute a specific subtask within the predefined fixed time period, which makes them inflexible and not suitable for diverse and dynamic scenarios with constantly changing subtasks. To break through above restrictions, a \textbf{S}liding \textbf{M}ultidimensional t\textbf{A}sk window based m\textbf{U}ti-agent reinforcement learnin\textbf{G} framework (SMAUG) is proposed for adaptive real-time subtask recognition. It leverages a sliding multidimensional task window to extract essential information of subtasks from trajectory segments concatenated based on observed and predicted trajectories in varying lengths. An inference network is designed to iteratively predict future trajectories with the subtask-oriented policy network. Furthermore, intrinsic motivation rewards are defined to promote subtask exploration and behavior diversity. SMAUG can be integrated with any Q-learning-based approach. Experiments on StarCraft II show that SMAUG not only demonstrates performance superiority in comparison with all baselines but also presents a more prominent and swift rise in rewards during the initial training stage.

cs.AI↗

arXiv2024

Learning to Retrieve for Job Matching

Web-scale search systems typically tackle the scalability challenge with a two-step paradigm: retrieval and ranking. The retrieval step, also known as candidate selection, often involves extracting standardized entities, creating an inverted index, and performing term matching for retrieval. Such traditional methods require manual and time-consuming development of query models. In this paper, we discuss applying learning-to-retrieve technology to enhance LinkedIns job search and recommendation systems. In the realm of promoted jobs, the key objective is to improve the quality of applicants, thereby delivering value to recruiter customers. To achieve this, we leverage confirmed hire data to construct a graph that evaluates a seeker's qualification for a job, and utilize learned links for retrieval. Our learned model is easy to explain, debug, and adjust. On the other hand, the focus for organic jobs is to optimize seeker engagement. We accomplished this by training embeddings for personalized retrieval, fortified by a set of rules derived from the categorization of member feedback. In addition to a solution based on a conventional inverted index, we developed an on-GPU solution capable of supporting both KNN and term matching efficiently.

cs.IR↗

arXiv2024

LinkSAGE: Optimizing Job Matching Using Graph Neural Networks

We present LinkSAGE, an innovative framework that integrates Graph Neural Networks (GNNs) into large-scale personalized job matching systems, designed to address the complex dynamics of LinkedIns extensive professional network. Our approach capitalizes on a novel job marketplace graph, the largest and most intricate of its kind in industry, with billions of nodes and edges. This graph is not merely extensive but also richly detailed, encompassing member and job nodes along with key attributes, thus creating an expansive and interwoven network. A key innovation in LinkSAGE is its training and serving methodology, which effectively combines inductive graph learning on a heterogeneous, evolving graph with an encoder-decoder GNN model. This methodology decouples the training of the GNN model from that of existing Deep Neural Nets (DNN) models, eliminating the need for frequent GNN retraining while maintaining up-to-date graph signals in near realtime, allowing for the effective integration of GNN insights through transfer learning. The subsequent nearline inference system serves the GNN encoder within a real-world setting, significantly reducing online latency and obviating the need for costly real-time GNN infrastructure. Validated across multiple online A/B tests in diverse product scenarios, LinkSAGE demonstrates marked improvements in member engagement, relevance matching, and member retention, confirming its generalizability and practical impact.

cs.LG↗

arXiv2023

Engineering of energy band and its impact on light transmission in non-reciprocal Hermitian hourglass lattice

We study a quasi-one-dimensional non-reciprocal Hermitian hourglass photonic lattice that can accomplish multiple functions. Under the effect of non-reciprocal coupling, this lattice can produce an energy isolation effect, two kinds of flat bands, and energy band inversion. The excitation and propagation of a single energy band and multiple energy bands can be realized; in the flat band condition, the system has compact localized states, and the flat bands can be excited by a straightforward method. In addition, we investigate the edge states under the open boundary condition; a double edge state appears by using a defect in the system. Our findings advance the theory of energy band regulation in artificial photonic lattices.

physics.optics↗

arXiv2023

RECESS Vaccine for Federated Learning: Proactive Defense Against Model Poisoning Attacks

Model poisoning attacks greatly jeopardize the application of federated learning (FL). The effectiveness of existing defenses is susceptible to the latest model poisoning attacks, leading to a decrease in prediction accuracy. Besides, these defenses are intractable to distinguish benign outliers from malicious gradients, which further compromises the model generalization. In this work, we propose a novel proactive defense named RECESS against model poisoning attacks. Different from the passive analysis in previous defenses, RECESS proactively queries each participating client with a delicately constructed aggregation gradient, accompanied by the detection of malicious clients according to their responses with higher accuracy. Furthermore, RECESS uses a new trust scoring mechanism to robustly aggregate gradients. Unlike previous methods that score each iteration, RECESS considers clients' performance correlation across multiple iterations to estimate the trust score, substantially increasing fault tolerance. Finally, we extensively evaluate RECESS on typical model architectures and four datasets under various settings. We also evaluated the defensive effectiveness against other types of poisoning attacks, the sensitivity of hyperparameters, and adaptive adversarial attacks. Experimental results show the superiority of RECESS in terms of reducing accuracy loss caused by the latest model poisoning attacks over five classic and two state-of-the-art defenses.

cs.CR↗

arXiv2023

Optimization of Image Transmission in a Cooperative Semantic Communication Networks

In this paper, a semantic communication framework for image transmission is developed. In the investigated framework, a set of servers cooperatively transmit images to a set of users utilizing semantic communication techniques. To evaluate the performance of studied semantic communication system, a multimodal metric is proposed to measure the correlation between the extracted semantic information and the original image. To meet the ISS requirement of each user, each server must jointly determine the semantic information to be transmitted and the resource blocks (RBs) used for semantic information transmission. We formulate this problem as an optimization problem aiming to minimize each server's transmission latency while reaching the ISS requirement. To solve this problem, a value decomposition based entropy-maximized multi-agent reinforcement learning (RL) is proposed, which enables servers to coordinate for training and execute RB allocation in a distributed manner to approach to a globally optimal performance with less training iterations. Compared to traditional multi-agent RL, the proposed RL improves the valuable action exploration of servers and the probability of finding a globally optimal RB allocation policy based on local observation. Simulation results show that the proposed algorithm can reduce the transmission delay by up to 16.1% compared to traditional multi-agent RL.

cs.AI↗

arXiv2022

Acoustic SLAM based on the Direction-of-Arrival and the Direct-to-Reverberant Energy Ratio

This paper proposes a new method that fuses acoustic measurements in the reverberation field and low-accuracy inertial measurement unit (IMU) motion reports for simultaneous localization and mapping (SLAM). Different from existing studies that only use acoustic data for direction-of-arrival (DoA) estimates, the source's distance from sensors is calculated with the direct-to-reverberant energy ratio (DRR) and applied as a new constraint to eliminate the nonlinear noise from motion reports. A particle filter is applied to estimate the critical distance, which is key for associating the source's distance with the DRR. A keyframe method is used to eliminate the deviation of the source position estimation toward the robot. The proposed DoA-DRR acoustic SLAM (D-D SLAM) is designed for three-dimensional motion and is suitable for most robots. The method is the first acoustic SLAM algorithm that has been validated on a real-world indoor scene dataset that contains only acoustic data and IMU measurements. Compared with previous methods, D-D SLAM has acceptable performance in locating the robot and building a source map from a real-world indoor dataset. The average location accuracy is 0.48 m, while the source position error converges to less than 0.25 m within 2.8 s. These results prove the effectiveness of D-D SLAM in real-world indoor scenes, which may be especially useful in search and rescue missions after disasters where the environment is foggy, i.e., unsuitable for light or laser irradiation.

cs.RO↗