arXiv subjects

Yuqi Zhou

Publications and source records attributed to Yuqi Zhou.

At least 37 records · Page 2Linked to original sources

arXiv2025

Length-Induced Embedding Collapse in PLM-based Models

Text embeddings from PLM-based models enable a wide range of applications, yet their performance often degrades on longer texts. In this paper, we introduce a phenomenon we call Length Collapse, where embeddings of longer texts tend to cluster together. This clustering results in a distributional inconsistency between the embeddings of short and long texts. We further investigate how these differences contribute to the performance decline observed with longer texts across various downstream tasks. Through a rigorous theoretical analysis of the self-attention mechanism, which acts as a low-pass filter in PLM-based models, we demonstrate that as text length increases, the strength of low-pass filtering intensifies, causing embeddings to retain more low-frequency components. As a result, input token features become more similar, leading to clustering and ultimately the collapse of embeddings for longer texts. To address this issue, we propose a simple method, TempScale, which mitigates the Length Collapse phenomenon. By narrowing the gap in low-pass filtering rates between long and short texts, TempScale ensures more consistent embeddings across different text lengths. This approach leads to performance improvements of 0.94% on MTEB and 1.10% on LongEmbed, which focuses specifically on long-context retrieval, providing strong evidence for the validity of our analysis. The source code is available at https://github.com/Yuqi-Zhou/Length_Collapse.

cs.CL↗

arXiv2025

GUI-G1: Understanding R1-Zero-Like Training for Visual Grounding in GUI Agents

Recent Graphical User Interface (GUI) agents replicate the R1-Zero paradigm, coupling online Reinforcement Learning (RL) with explicit chain-of-thought reasoning prior to object grounding and thereby achieving substantial performance gains. In this paper, we first conduct extensive analysis experiments of three key components of that training pipeline: input design, output evaluation, and policy update-each revealing distinct challenges arising from blindly applying general-purpose RL without adapting to GUI grounding tasks. Input design: Current templates encourage the model to generate chain-of-thought reasoning, but longer chains unexpectedly lead to worse grounding performance. Output evaluation: Reward functions based on hit signals or box area allow models to exploit box size, leading to reward hacking and poor localization quality. Policy update: Online RL tends to overfit easy examples due to biases in length and sample difficulty, leading to under-optimization on harder cases. To address these issues, we propose three targeted solutions. First, we adopt a Fast Thinking Template that encourages direct answer generation, reducing excessive reasoning during training. Second, we incorporate a box size constraint into the reward function to mitigate reward hacking. Third, we revise the RL objective by adjusting length normalization and adding a difficulty-aware scaling factor, enabling better optimization on hard samples. Our GUI-G1-3B, trained on 17K public samples with Qwen2.5-VL-3B-Instruct, achieves 90.3% accuracy on ScreenSpot and 37.1% on ScreenSpot-Pro. This surpasses all prior models of similar size and even outperforms the larger UI-TARS-7B, establishing a new state-of-the-art in GUI agent grounding. The project repository is available at https://github.com/Yuqi-Zhou/GUI-G1.

cs.CL↗

arXiv2025

Controlling Avatar Diffusion with Learnable Gaussian Embedding

Recent advances in diffusion models have made significant progress in digital human generation. However, most existing models still struggle to maintain 3D consistency, temporal coherence, and motion accuracy. A key reason for these shortcomings is the limited representation ability of commonly used control signals(e.g., landmarks, depth maps, etc.). In addition, the lack of diversity in identity and pose variations in public datasets further hinders progress in this area. In this paper, we analyze the shortcomings of current control signals and introduce a novel control signal representation that is optimizable, dense, expressive, and 3D consistent. Our method embeds a learnable neural Gaussian onto a parametric head surface, which greatly enhances the consistency and expressiveness of diffusion-based head models. Regarding the dataset, we synthesize a large-scale dataset with multiple poses and identities. In addition, we use real/synthetic labels to effectively distinguish real and synthetic data, minimizing the impact of imperfections in synthetic data on the generated head images. Extensive experiments show that our model outperforms existing methods in terms of realism, expressiveness, and 3D consistency. Our code, synthetic datasets, and pre-trained models will be released in our project page: https://ustc3dv.github.io/Learn2Control/

cs.GR↗

arXiv2025

CHOP: Mobile Operating Assistant with Constrained High-frequency Optimized Subtask Planning

The advancement of visual language models (VLMs) has enhanced mobile device operations, allowing simulated human-like actions to address user requirements. Current VLM-based mobile operating assistants can be structured into three levels: task, subtask, and action. The subtask level, linking high-level goals with low-level executable actions, is crucial for task completion but faces two challenges: ineffective subtasks that lower-level agent cannot execute and inefficient subtasks that fail to contribute to the completion of the higher-level task. These challenges stem from VLM's lack of experience in decomposing subtasks within GUI scenarios in multi-agent architecture. To address these, we propose a new mobile assistant architecture with constrained high-frequency o}ptimized planning (CHOP). Our approach overcomes the VLM's deficiency in GUI scenarios planning by using human-planned subtasks as the basis vector. We evaluate our architecture in both English and Chinese contexts across 20 Apps, demonstrating significant improvements in both effectiveness and efficiency. Our dataset and code is available at https://github.com/Yuqi-Zhou/CHOP

cs.AI↗

arXiv2025

Machine Learning for Fairness-Aware Load Shedding: A Real-Time Solution via Identifying Binding Constraints

Timely and effective load shedding in power systems is critical for maintaining supply-demand balance and preventing cascading blackouts. To eliminate load shedding bias against specific regions in the system, optimization-based methods are uniquely positioned to help balance between economic and fairness considerations. However, the resulting optimization problem involves complex constraints, which can be time-consuming to solve and thus cannot meet the real-time requirements of load shedding. To tackle this challenge, in this paper we present an efficient machine learning algorithm to enable millisecond-level computation for the optimization-based load shedding problem. Numerical studies on both a 3-bus toy example and a realistic RTS-GMLC system have demonstrated the validity and efficiency of the proposed algorithm for delivering fairness-aware and real-time load shedding decisions.

eess.SY↗

arXiv2025

GUI Agents with Foundation Models: A Comprehensive Survey

Recent advances in foundation models, particularly Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs), have facilitated the development of intelligent agents capable of performing complex tasks. By leveraging the ability of (M)LLMs to process and interpret Graphical User Interfaces (GUIs), these agents can autonomously execute user instructions, simulating human-like interactions such as clicking and typing. This survey consolidates recent research on (M)LLM-based GUI agents, highlighting key innovations in data resources, frameworks, and applications. We begin by reviewing representative datasets and benchmarks, followed by an overview of a generalized, unified framework that encapsulates the essential components of prior studies, supported by a detailed taxonomy. Additionally, we explore relevant commercial applications. Drawing insights from existing work, we identify key challenges and propose future research directions. We hope this survey will inspire further advancements in the field of (M)LLM-based GUI agents.

cs.AI↗

arXiv2025

Machine Learning for Scalable and Optimal Load Shedding Under Power System Contingency

Prompt and effective corrective actions in response to unexpected contingencies are crucial for improving power system resilience and preventing cascading blackouts. The optimal load shedding (OLS) accounting for network limits has the potential to address the diverse system-wide impacts of contingency scenarios as compared to traditional local schemes. However, due to the fast cascading propagation of initial contingencies, real-time OLS solutions are challenging to attain in large systems with high computation and communication needs. In this paper, we propose a decentralized design that leverages offline training of a neural network (NN) model for individual load centers to autonomously construct the OLS solutions from locally available measurements. Our learning-for-OLS approach can greatly reduce the computation and communication needs during online emergency responses, thus preventing the cascading propagation of contingencies for enhanced power grid resilience. Numerical studies on both the IEEE 118-bus system and a synthetic Texas 2000-bus system have demonstrated the efficiency and effectiveness of our scalable OLS learning design for timely power system emergency operations.

cs.LG↗

arXiv2025

Exploring the Use of Autonomous Unmanned Vehicles for Supporting Power Grid Operations

This paper explores the use of autonomous unmanned vehicles to support power grid operations. With built-in batteries and the capability to carry additional battery energy storage, the rising number of autonomous vehicles can represent a substantial amount of capacity that is currently underutilized in the power grid. Unlike traditional electric vehicles that require drivers, the operations of autonomous vehicles can be performed without human intervention. To guide idle vehicles to autonomously support power grids, we propose a tractable optimization-based method to effectively integrate these "mobile batteries" into grid operations. During real-time operations, the vehicles are strategically routed to target locations to maintain power balance and reduce operating costs. Numerical studies have confirmed both the validity and the scalability of the proposed algorithm to efficiently integrate autonomous vehicles into routine power system operations.

eess.SY↗

arXiv2024

Programmable simulation of high-order exceptional point with a trapped ion

The nontrivial degeneracies in non-Hermitian systems, exceptional points (EPs), have attracted extensive attention due to intriguing phenomena. Compared with commonly observed second-order EPs, high-order EPs show rich physics due to their extended dimension and parameter space, ranging from the coalescence of EPs into higher order to potential applications in topological properties. However, these features also pose challenges in controlling multiple coherent and dissipative elements in a scaled system. Here we experimentally demonstrate a native programmable control to simulate a high-order non-Hermitian Hamiltonian in a multi-dimensional trapped ion system. We simulate a series of non-Hermitian systems with varied parameters and observe the coalescence of second-order EPs into a fourth-order EP. Our results pave the way for scalable quantum simulation of high-dimensional dissipative systems and can be beneficial for the application of high-order EPs in quantum sensing and quantum control.

quant-ph↗

arXiv2024

Machine Learning-Assisted Distribution System Network Reconfiguration Problem

High penetration from volatile renewable energy resources in the grid and the varying nature of loads raise the need for frequent line switching to ensure the efficient operation of electrical distribution networks. Operators must ensure maximum load delivery, reduced losses, and the operation between voltage limits. However, computations to decide the optimal feeder configuration are often computationally expensive and intractable, making it unfavorable for real-time operations. This is mainly due to the existence of binary variables in the network reconfiguration optimization problem. To tackle this issue, we have devised an approach that leverages machine learning techniques to reshape distribution networks featuring multiple substations. This involves predicting the substation responsible for serving each part of the network. Hence, it leaves simple and more tractable Optimal Power Flow problems to be solved. This method can produce accurate results in a significantly faster time, as demonstrated using the IEEE 37-bus distribution feeder. Compared to the traditional optimization-based approaches, a feasible solution is achieved approximately ten times faster for all the tested scenarios.

eess.SY↗

arXiv2024

Transmission Line Outage Probability Prediction Under Extreme Events Using Peter-Clark Bayesian Structural Learning

Recent years have seen a notable increase in the frequency and intensity of extreme weather events. With a rising number of power outages caused by these events, accurate prediction of power line outages is essential for safe and reliable operation of power grids. The Bayesian network is a probabilistic model that is very effective for predicting line outages under weather-related uncertainties. However, most existing studies in this area offer general risk assessments, but fall short of providing specific outage probabilities. In this work, we introduce a novel approach for predicting transmission line outage probabilities using a Bayesian network combined with Peter-Clark (PC) structural learning. Our approach not only enables precise outage probability calculations, but also demonstrates better scalability and robust performance, even with limited data. Case studies using data from BPA and NOAA show the effectiveness of this approach, while comparisons with several existing methods further highlight its advantages.

cs.LG↗

arXiv2024

Equitable Networked Microgrid Topology Reconfiguration for Wildfire Risk Mitigation

The increasing number of wildfires in recent years consistently challenges the safe and reliable operations of power systems. To prevent power lines and other electrical components from causing wildfires under extreme conditions, electric utilities often deploy public safety power shutoffs (PSPS) to mitigate the wildfire risks therein. Although PSPS are effective countermeasures against wildfires, uncoordinated strategies can cause disruptions in electricity supply and even lead to cascading failures. Meanwhile, it is important to consider mitigating biased decisions on different communities and populations during the implementation of shutoff actions. In this work, we primarily focus on the dynamic reconfiguration problem of networked microgrids with distributed energy resources. In particular, we formulate a rolling horizon optimization problem allowing for flexible network reconfiguration at each time interval to mitigate wildfire risks. To promote equity and fairness during the span of shutoffs, we further enforce a range of constraints associated with load shedding to discourage disproportionate impact on individual load blocks. Numerical studies on a modified IEEE 13-bus system and a larger-sized Smart-DS system demonstrate the performance of the proposed algorithm towards more equitable power shutoff operations.

eess.SY↗

arXiv2024

Mitigating the Impact of Uncertain Wildfire Risk on Power Grids through Topology Control

Wildfires pose a significant threat to the safe and reliable operations of the electric grid. To mitigate wildfire risk, system operators resort to public safety power shutoffs, or PSPS, that shed load for a subset of customers. As wildfire risk forecasts are stochastic, such decision-making may often be sub-optimal. This paper proposes a two-stage topology control problem that jointly minimizes generation and load-shedding costs in the face of uncertain fire risk. Compared to existing work, we include preand post-event topology control actions and consider scenarios where the wildfire risk is known with low and high confidence. The effectiveness of the proposed approach is demonstrated using a benchmark test system, artificially geo-located in Southern California, and using stochastic wildfire risk data that exists in the literature. Our work provides a crucial study of the comparative benefits of pre-event versus post-event control and the effects of wildfire risk accuracy on each control strategy.

eess.SY↗

arXiv2024

Neural Retrievers are Biased Towards LLM-Generated Content

Recently, the emergence of large language models (LLMs) has revolutionized the paradigm of information retrieval (IR) applications, especially in web search, by generating vast amounts of human-like texts on the Internet. As a result, IR systems in the LLM era are facing a new challenge: the indexed documents are now not only written by human beings but also automatically generated by the LLMs. How these LLM-generated documents influence the IR systems is a pressing and still unexplored question. In this work, we conduct a quantitative evaluation of IR models in scenarios where both human-written and LLM-generated texts are involved. Surprisingly, our findings indicate that neural retrieval models tend to rank LLM-generated documents higher. We refer to this category of biases in neural retrievers towards the LLM-generated content as the \textbf{source bias}. Moreover, we discover that this bias is not confined to the first-stage neural retrievers, but extends to the second-stage neural re-rankers. Then, in-depth analyses from the perspective of text compression indicate that LLM-generated texts exhibit more focused semantics with less noise, making it easier for neural retrieval models to semantic match. To mitigate the source bias, we also propose a plug-and-play debiased constraint for the optimization objective, and experimental results show its effectiveness. Finally, we discuss the potential severe concerns stemming from the observed source bias and hope our findings can serve as a critical wake-up call to the IR community and beyond. To facilitate future explorations of IR in the LLM era, the constructed two new benchmarks are available at https://github.com/KID-22/Source-Bias.

cs.IR↗

arXiv2024

Virtual Context: Enhancing Jailbreak Attacks with Special Token Injection

Jailbreak attacks on large language models (LLMs) involve inducing these models to generate harmful content that violates ethics or laws, posing a significant threat to LLM security. Current jailbreak attacks face two main challenges: low success rates due to defensive measures and high resource requirements for crafting specific prompts. This paper introduces Virtual Context, which leverages special tokens, previously overlooked in LLM security, to improve jailbreak attacks. Virtual Context addresses these challenges by significantly increasing the success rates of existing jailbreak methods and requiring minimal background knowledge about the target model, thus enhancing effectiveness in black-box settings without additional overhead. Comprehensive evaluations show that Virtual Context-assisted jailbreak attacks can improve the success rates of four widely used jailbreak methods by approximately 40% across various LLMs. Additionally, applying Virtual Context to original malicious behaviors still achieves a notable jailbreak effect. In summary, our research highlights the potential of special tokens in jailbreak attacks and recommends including this threat in red-teaming testing to comprehensively enhance LLM security.

cs.CR↗

arXiv2024

Cocktail: A Comprehensive Information Retrieval Benchmark with LLM-Generated Documents Integration

The proliferation of Large Language Models (LLMs) has led to an influx of AI-generated content (AIGC) on the internet, transforming the corpus of Information Retrieval (IR) systems from solely human-written to a coexistence with LLM-generated content. The impact of this surge in AIGC on IR systems remains an open question, with the primary challenge being the lack of a dedicated benchmark for researchers. In this paper, we introduce Cocktail, a comprehensive benchmark tailored for evaluating IR models in this mixed-sourced data landscape of the LLM era. Cocktail consists of 16 diverse datasets with mixed human-written and LLM-generated corpora across various text retrieval tasks and domains. Additionally, to avoid the potential bias from previously included dataset information in LLMs, we also introduce an up-to-date dataset, named NQ-UTD, with queries derived from recent events. Through conducting over 1,000 experiments to assess state-of-the-art retrieval models against the benchmarked datasets in Cocktail, we uncover a clear trade-off between ranking performance and source bias in neural retrieval models, highlighting the necessity for a balanced approach in designing future IR systems. We hope Cocktail can serve as a foundational resource for IR research in the LLM era, with all data and code publicly available at \url{https://github.com/KID-22/Cocktail}.

cs.IR↗

arXiv2024

SL(n) Contravariant Matrix-Valued Valuations on Polytopes

All $\textrm{SL}(n)$ contravariant matrix-valued valuations on polytopes in $\mathbb{R}^n$ are completely classified without any continuity assumptions. Moreover, the symmetry assumption of matrices is removed. The general Lutwak-Yang-Zhang matrix turns out to be the only such valuation if $n\geq 4$, while a new function shows up in dimension three. In dimension two, the classification corresponds to the known case of $\textrm{SL}(2)$ equivariant matrix-valued valuations.

math.DG↗

arXiv2023

Some Sharp Chernoff type inequalities

Two sharp Chernoff type inequalities are obtained for star body in $\mathbb{R}^2$, one of which is an extension of the dual Chernoff-Ou-Pan inequality, and the other is the reverse Chernoff type inequality. Furthermore, we establish a generalized dual symmetric mixed Chernoff inequality for two planar star bodies. As a direct consequence, a new proof of the the dual symmetric mixed isoperimetric inequality is presented.

math.DG↗