arXiv subjects

Yingchen Zhang

Publications and source records attributed to Yingchen Zhang.

17 recordsLinked to original sources

arXiv2026

Token-Flow Firewall: Semantic Runtime Auditing for Persistent AI Agents

Persistent AI agents extend large language models (LLMs) beyond single-turn interaction into long-lived software systems. Unlike traditional chat assistants, unsafe content in these agents can propagate through persistent state, reusable skills, and tool-mediated interactions, creating a substantially larger semantic attack surface. We observe that most security-critical interactions in such agents are transmitted through natural-language token flows, including memory updates, tool arguments, retrieved files, and inter-component communications. This observation enables a new security formulation: unsafe behavior can be intercepted as risky semantic flows before reaching privileged runtime sinks. Based on this insight, we propose TokenWall, a runtime defense framework that acts as a semantic firewall over agent token flows. TokenWall performs boundary-aware semantic auditing over these flows, constructing structured source-sink audit records, applying lightweight local inspection before execution, and selectively escalating ambiguous high-risk cases to stronger arbitration modules. Unlike prior approaches that rely on sparse auditing or remote large-model oversight, TokenWall enables full-coverage pre-execution mediation while reducing remote arbitration and latency. Experiments on CIK-Bench show that TokenWall reduces attack success rate to 12.5% while maintaining a 97.4% benign executable pass rate without human confirmation. TokenWall further introduces only 0.69 seconds of additional latency on benign cases, demonstrating that semantic runtime containment can achieve a practical security-utility trade-off for persistent AI agents.

cs.CR↗

arXiv2026

G-DRAGON: Geospatial Reasoning and Dynamic Planning for Retrieval-Augmented Outdoor Navigation

Autonomous ground robots operating in large-scale outdoor environments require both robust long-range navigation and fine-grained ''last-mile'' exploration. Current advances in visual-language navigation (VLN) work well at short-range tasks, lacking geospatial grounding for long-distance missions. Some OpenStreetMap (OSM)-based methods relying on cloud-based Large Language Models (LLMs) are prone to factual hallucination and cannot conduct ''last-mile'' exploration based on human instruction. To address these challenges, we present G-DRAGON, a retrieval-augmented framework for outdoor, open-world navigation. This framework maps natural-language commands to versioned, local OSM entities via generative retrieval based on lightweight LLM, yielding accurate coordinates for global route planning. A high-level planning module bridges global topological routes with the SLAM system, projecting geospatial waypoints into the robot's navigable frame. For the ''last mile," the framework transitions to frontier-based exploration and open-set semantic voxel mapping to localize open-vocabulary targets. Experimental results in simulation demonstrate our framework outperforms state-of-the-art baselines. Furthermore, we validate the system in unseen real-world urban environments on an Unmanned Ground Vehicle (UGV), successfully completing person-search missions with trajectories of up to 500m.

cs.RO↗

arXiv2025

C2T-ID: Converting Semantic Codebooks to Textual Document Identifiers for Generative Search

Designing document identifiers (docids) that carry rich semantic information while maintaining tractable search spaces is a important challenge in generative retrieval (GR). Popular codebook methods address this by building a hierarchical semantic tree and constraining generation to its child nodes, yet their numeric identifiers cannot leverage the large language model's pretrained natural language understanding. Conversely, using text as docid provides more semantic expressivity but inflates the decoding space, making the system brittle to early-step errors. To resolve this trade-off, we propose C2T-ID: (i) first construct semantic numerical docid via hierarchical clustering; (ii) then extract high-frequency metadata keywords and iteratively replace each numeric label with its cluster's top-K keywords; and (iii) an optional two-level semantic smoothing step further enhances the fluency of C2T-ID. Experiments on Natural Questions and Taobao's product search demonstrate that C2T-ID significantly outperforms atomic, semantic codebook, and pure-text docid baselines, demonstrating its effectiveness in balancing semantic expressiveness with search space constraints.

cs.IR↗

arXiv2025

Retrieval-in-the-Chain: Bootstrapping Large Language Models for Generative Retrieval

Generative retrieval (GR) is an emerging paradigm that leverages large language models (LLMs) to autoregressively generate document identifiers (docids) relevant to a given query. Prior works have focused on leveraging the generative capabilities of LLMs to improve GR, while overlooking that their reasoning capabilities could likewise help. This raises a key question: Can explicit reasoning benefit GR? To investigate, we first conduct a preliminary study where an LLM is prompted to generate free-form chain-of-thought (CoT) reasoning before performing constrained docid decoding. Although this method outperforms standard GR, the generated reasoning tends to be verbose and poorly aligned with the docid space. These limitations motivate the development of a reasoning mechanism better tailored to GR. Therefore, we propose Reason-for-Retrieval (R4R), a reasoning-augmented framework for GR that converts free-form CoT reasoning into a compact, structured format, and iteratively refines the reasoning during the retrieval process. R4R augments an existing GR method by leveraging a reasoning-capable LLM that has been instruction-tuned for GR. At inference time, R4R first uses the LLM to generate an initial structured reasoning; then the same LLM alternates between (i) constrained decoding with the chosen GR method to produce candidate docids and (ii) updating the reasoning based on retrieval results to improve the next round. R4R does not require additional models or training, and instead a single LLM serves as both the reasoning generator and the retriever. Extensive experiments on Natural Questions, MS MARCO, and a real-world item-search benchmark validate the effectiveness of R4R.

cs.IR↗

arXiv2025

Does Generative Retrieval Overcome the Limitations of Dense Retrieval?

Generative retrieval (GR) has emerged as a new paradigm in neural information retrieval, offering an alternative to dense retrieval (DR) by directly generating identifiers of relevant documents. In this paper, we theoretically and empirically investigate how GR fundamentally diverges from DR in both learning objectives and representational capacity. GR performs globally normalized maximum-likelihood optimization and encodes corpus and relevance information directly in the model parameters, whereas DR adopts locally normalized objectives and represents the corpus with external embeddings before computing similarity via a bilinear interaction. Our analysis suggests that, under scaling, GR can overcome the inherent limitations of DR, yielding two major benefits. First, with larger corpora, GR avoids the sharp performance degradation caused by the optimization drift induced by DR's local normalization. Second, with larger models, GR's representational capacity scales with parameter size, unconstrained by the global low-rank structure that limits DR. We validate these theoretical insights through controlled experiments on the Natural Questions and MS MARCO datasets, across varying negative sampling strategies, embedding dimensions, and model scales. But despite its theoretical advantages, GR does not universally outperform DR in practice. We outline directions to bridge the gap between GR's theoretical potential and practical performance, providing guidance for future research in scalable and robust generative retrieval.

cs.IR↗

arXiv2022

Physics-Informed Statistical Modeling for Wildfire Aerosols Process Using Multi-Source Geostationary Satellite Remote-Sensing Data Streams

Increasingly frequent wildfires significantly affect solar energy production as the atmospheric aerosols generated by wildfires diminish the incoming solar radiation to the earth. Atmospheric aerosols are measured by Aerosol Optical Depth (AOD), and AOD data streams can be retrieved and monitored by geostationary satellites. However, multi-source remote-sensing data streams often present heterogeneous characteristics, including different data missing rates, measurement errors, systematic biases, and so on. To accurately estimate and predict the underlying AOD propagation process, there exist practical needs and theoretical interests to propose a physics-informed statistical approach for modeling wildfire AOD propagation by simultaneously utilizing, or fusing, multi-source heterogeneous satellite remote-sensing data streams. Leveraging a spectral approach, the proposed approach integrates multi-source satellite data streams with a fundamental advection-diffusion equation that governs the AOD propagation process. A bias correction process is included in the statistical model to account for the bias of the physics model and the truncation error of the Fourier series. The proposed approach is applied to California wildfires AOD data streams obtained from the National Oceanic and Atmospheric Administration. Comprehensive numerical examples are provided to demonstrate the predictive capabilities and model interpretability of the proposed approach. Computer code has been made available on GitHub.

stat.AP↗

arXiv2021

Measurement placement in electric power transmission and distribution grids: Review of concepts, methods, and research needs

Sensing and measurement systems are quintessential to the safe and reliable operation of electric power grids. Their strategic placement is of ultimate importance because it is not economically viable to install measurement systems on every node and branch of a power grid, though they need to be monitored. An overwhelming number of strategies have been developed to meet oftentimes multiple conflicting objectives. The prime challenge in formulating the problem lies in developing a heuristic or an optimization model that, though mathematically tractable and constrained in cost, leads to trustworthy technical solutions. Further, large-scale, long-term deployments pose additional challenges because the boundary conditions change as technologies evolve. For instance, the advent of new technologies in sensing and measurement, as well as in communications and networking, might impact the cost and performance of available solutions and shift initially set conditions. Also, the placement strategies developed for transmission grids might not be suitable for distribution grids, and vice versa, because of unique characteristics; therefore, the strategies need to be flexible, to a certain extent, because no two power grids are alike. Despite the extensive literature on the present topic, the focus of published works tends to be on a specific subject, such as the optimal placement of measurements to ensure observability in transmission grids. There is a dearth of work providing a comprehensive picture for developing optimal placement strategies. Because of the ongoing efforts on the modernization of electric power grids, there is a need to consolidate the status quo while exposing its limitations to inform policymakers, industry stakeholders, and researchers on the research-and-development needs to push the boundaries for innovation.

eess.SY↗

arXiv2021

On analytical construction of observable functions in extended dynamic mode decomposition for nonlinear estimation and prediction

We propose an analytical construction of observable functions in the extended dynamic mode decomposition (EDMD) algorithm. EDMD is a numerical method for approximating the spectral properties of the Koopman operator. The choice of observable functions is fundamental for the application of EDMD to nonlinear problems arising in systems and control. Existing methods either start from a set of dictionary functions and look for the subset that best fits the underlying nonlinear dynamics or they rely on machine learning algorithms to "learn" observable functions. Conversely, in this paper, we start from the dynamical system model and lift it through the Lie derivatives, rendering it into a polynomial form. This proposed transformation into a polynomial form is exact, and it provides an adequate set of observable functions. The strength of the proposed approach is its applicability to a broader class of nonlinear dynamical systems, particularly those with nonpolynomial functions and compositions thereof. Moreover, it retains the physical interpretability of the underlying dynamical system and can be readily integrated into existing numerical libraries. The proposed approach is illustrated with an application to electric power systems. The modeled system consists of a single generator connected to an infinite bus, where nonlinear terms include sine and cosine functions. The results demonstrate the effectiveness of the proposed procedure in off-attractor nonlinear dynamics for estimation and prediction; the observable functions obtained from the proposed construction outperform methods that use dictionary functions comprising monomials or radial basis functions.

eess.SY↗

arXiv2020

Matrix Completion Using Alternating Minimization for Distribution System State Estimation

This paper examines the problem of state estimation in power distribution systems under low-observability conditions. The recently proposed constrained matrix completion method which combines the standard matrix completion method and power flow constraints has been shown to be effective in estimating voltage phasors under low-observability conditions using single-snapshot information. However, the method requires solving a semidefinite programming (SDP) problem, which becomes computationally infeasible for large systems and if multiple-snapshot (time-series) information is used. This paper proposes an efficient algorithm to solve the constrained matrix completion problem with time-series data. This algorithm is based on reformulating the matrix completion problem as a bilinear (non-convex) optimization problem, and applying the alternating minimization algorithm to solve this problem. This paper proves the summable convergence of the proposed algorithm, and demonstrates its efficacy and scalability via IEEE 123-bus system and a real utility feeder system. This paper also explores the value of adding more data from the history in terms of computation time and estimation accuracy.

math.OC↗

arXiv2020

Nonlinear Virtual Inertia Control of WTGs for Enhancing Primary Frequency Response and Suppressing Drive-Train Torsional Oscillations

Virtual inertia controllers (VICs) for wind turbine generators (WTGs) have been recently developed to compensate for the reduction of inertia in power systems. However, VICs can induce low-frequency torsional oscillations of the drive train of WTGs. This paper addresses this issue and develops a new nonlinear VIC based on objective holographic feedbacks theory. This approach allows transforming the objectives that require improvement into a completely controllable system of Brunovsky's type. Simulation results under various scenarios demonstrate that the proposed method outperforms existing VICs in terms of suppression of WTG low-frequency drive-train torsional oscillations, enhancement of system frequency nadir as well as fast and smooth recovery of WTG rotor speed to the original MPP before the disturbance. The proposed method is also able to coordinate multiple WTGs.

eess.SY↗

arXiv2020

Deep Reinforcement Learning Based Volt-VAR Optimization in Smart Distribution Systems

This paper develops a model-free volt-VAR optimization (VVO) algorithm via multi-agent deep reinforcement learning (MADRL) in unbalanced distribution systems. This method is novel since we cast the VVO problem in unbalanced distribution networks to an intelligent deep Q-network (DQN) framework, which avoids solving a specific optimization model directly when facing time-varying operating conditions of the systems. We consider statuses/ratios of switchable capacitors, voltage regulators, and smart inverters installed at distributed generators as the action variables of the DQN agents. A delicately designed reward function guides these agents to interact with the distribution system, in the direction of reinforcing voltage regulation and power loss reduction simultaneously. The forward-backward sweep method for radial three-phase distribution systems provides accurate power flow results within a few iterations to the DQN environment. Finally, the proposed multi-objective MADRL method realizes the dual goals for VVO. We test this algorithm on the unbalanced IEEE 13-bus and 123-bus systems. Numerical simulations validate the excellent performance of this method in voltage regulation and power loss reduction.

eess.SY↗

arXiv2020

A Data-Driven Game-Theoretic Approach for Behind-the-Meter PV Generation Disaggregation

Rooftop solar photovoltaic (PV) power generator is a widely used distributed energy resource (DER) in distribution systems. Currently, the majority of PVs are installed behind-the-meter (BTM), where only customers' net demand is recorded by smart meters. Disaggregating BTM PV generation from net demand is critical to utilities for enhancing grid-edge observability. In this paper, a data-driven approach is proposed for BTM PV generation disaggregation using solar and demand exemplars. First, a data clustering procedure is developed to construct a library of candidate load/solar exemplars. To handle the volatility of BTM resources, a novel game-theoretic learning process is proposed to adaptively generate optimal composite exemplars using the constructed library of candidate exemplars, through repeated evaluation of disaggregation residuals. Finally, the composite native demand and solar exemplars are employed to disaggregate solar generation from net demand using a semi-supervised source separator. The proposed methodology has been verified using real smart meter data and feeder models.

eess.SP↗

arXiv2019

Robust Matrix Completion State Estimation in Distribution Systems

Due to the insufficient measurements in the distribution system state estimation (DSSE), full observability and redundant measurements are difficult to achieve without using the pseudo measurements. The matrix completion state estimation (MCSE) combines the matrix completion and power system model to estimate voltage by exploring the low-rank characteristics of the matrix. This paper proposes a robust matrix completion state estimation (RMCSE) to estimate the voltage in a distribution system under a low-observability condition. Tradition state estimation weighted least squares (WLS) method requires full observability to calculate the states and needs redundant measurements to proceed a bad data detection. The proposed method improves the robustness of the MCSE to bad data by minimizing the rank of the matrix and measurements residual with different weights. It can estimate the system state in a low-observability system and has robust estimates without the bad data detection process in the face of multiple bad data. The method is numerically evaluated on the IEEE 33-node radial distribution system. The estimation performance and robustness of RMCSE are compared with the WLS with the largest normalized residual bad data identification (WLS-LNR), and the MCSE.

cs.LG↗

arXiv2019

Matrix Completion for Low-Observability Voltage Estimation

With the rising penetration of distributed energy resources, distribution system control and enabling techniques such as state estimation have become essential to distribution system operation. However, traditional state estimation techniques have difficulty coping with the low-observability conditions often present on the distribution system due to the paucity of sensors and heterogeneity of measurements. To address these limitations, we propose a distribution system state estimation algorithm that employs matrix completion (a tool for estimating missing values in low-rank matrices) augmented with noise-resilient power flow constraints. This method operates under low-observability conditions where standard least-squares-based methods cannot operate, and flexibly incorporates any network quantities measured in the field. We empirically evaluate our method on the IEEE 33- and 123-bus test systems, and find that it provides near-perfect state estimation performance (within 1\% mean absolute percent error) across many low-observability data availability regimes.

math.OC↗

arXiv2017

Power-Traffic Coordinated Operation for Bi-Peak Shaving and Bi-Ramp Smoothing -A Hierarchical Data-Driven Approach

With the rapid adoption of distributed photovoltaics (PVs) in certain regions, issues such as lower net load valley during the day and more steep ramping of the demand after sunset start to challenge normal operations at utility companies. Urban transportation systems also have high peak congestion periods and steep ramping because of traffic patterns. We propose using the emerging electric vehicles (EVs) and the charing/discharging stations (CDSs) to coordinate the operation between power distribution system (PDS) and the urban transportation system (UTS), therefore, the operation challenges in each system can be mitigated by utilizing the flexibility of the other system. We conducted the simulation and numerical analysis using the IEEE 8,500-bus for the PDS and the Sioux Falls system with about 10,000 cars for the UTS. Two systems are simulated jointly to demonstrate the feasibility and effectiveness of the proposed approach.

eess.SY↗

arXiv2017

Chance-Constrained Day-Ahead Hourly Scheduling in Distribution System Operation

This paper aims to propose a two-step approach for day-ahead hourly scheduling in a distribution system operation, which contains two operation costs, the operation cost at substation level and feeder level. In the first step, the objective is to minimize the electric power purchase from the day-ahead market with the stochastic optimization. The historical data of day-ahead hourly electric power consumption is used to provide the forecast results with the forecasting error, which is presented by a chance constraint and formulated into a deterministic form by Gaussian mixture model (GMM). In the second step, the objective is to minimize the system loss. Considering the nonconvexity of the three-phase balanced AC optimal power flow problem in distribution systems, the second-order cone program (SOCP) is used to relax the problem. Then, a distributed optimization approach is built based on the alternating direction method of multiplier (ADMM). The results shows that the validity and effectiveness method.

eess.SY↗

arXiv2017

Load Forecasting Based Distribution System Network Reconfiguration-A Distributed Data-Driven Approach

In this paper, a short-term load forecasting approach based network reconfiguration is proposed in a parallel manner. Specifically, a support vector regression (SVR) based short-term load forecasting approach is designed to provide an accurate load prediction and benefit the network reconfiguration. Because of the nonconvexity of the three-phase balanced optimal power flow, a second-order cone program (SOCP) based approach is used to relax the optimal power flow problem. Then, the alternating direction method of multipliers (ADMM) is used to compute the optimal power flow in distributed manner. Considering the limited number of the switches and the increasing computation capability, the proposed network reconfiguration is solved in a parallel way. The numerical results demonstrate the feasible and effectiveness of the proposed approach.

eess.SY↗