arXiv subjects

Yuchen Zheng

Publications and source records attributed to Yuchen Zheng.

17 recordsLinked to original sources

arXiv2026

FSGR: Mitigating Token Frequency Bias for Fair SID-Based Generative Recommendation

Semantic ID (SID)-based generative recommendation has recently achieved remarkable success. However, existing methods suffer from a previously overlooked fairness issue, which we term \textbf{Token Frequency Bias}, where high-frequency SID tokens are systematically over-predicted while low-frequency SID tokens are under-predicted. This bias originates from the combined effects of imbalanced semantic codebooks during SID construction, and popularity bias together with the maximum likelihood estimation objective during recommendation training, resulting in unfair exposure across item categories. Existing SID methods mainly focus on improving codebook quality and overlook the impact of token frequency imbalance on downstream recommendation fairness, while LLM debiasing methods often yield suboptimal results when directly applied to SID-based recommendation, due to the hierarchical semantics of SID tokens. To address this issue, we propose \textbf{FSGR}, a fairness optimization framework for SID-based generative recommendation. During SID construction, FSGR employs OT-based Assignment Optimization and Dual-Criteria Re-anchor mechanism to form a more balanced SID representation space. During recommendation training, it adopts a two-stage training strategy and introduces Hierarchical Frequency Calibration for layer-specific fairness fine-tuning. Experiments on three public datasets with three backbone models demonstrate that FSGR mitigates token frequency bias and delivers an average Gini fairness improvement of over 20\% while maintaining competitive recommendation accuracy.

cs.IR↗

arXiv2026

Symskill: Symbol and Skill Co-Invention for Data-Efficient and Reactive Long-Horizon Manipulation

Multi-step manipulation in dynamic environments remains challenging. Imitation learning (IL) is reactive but lacks compositional generalization, since monolithic policies do not decide which skill to reuse when scenes change. Classical task-and-motion planning (TAMP) offers compositionality, but its high planning latency prevents real-time failure recovery. We introduce SymSkill, a unified framework that jointly learns predicates, operators, and skills from unlabeled, unsegmented demonstrations, combining compositional generalization with real-time recovery. Offline, SymSkill learns symbolic abstractions and goal-oriented skills directly from demonstrations. Online, given a conjunction of learned predicates, it uses a symbolic planner to compose and reorder skills to achieve symbolic goals while recovering from failures at both the motion and symbolic levels in real time. Coupled with a compliant controller, SymSkill supports safe execution under human and environmental disturbances. In RoboCasa simulation, SymSkill executes 12 single-step tasks with 85% success and composes them into multi-step plans without additional data. On a real Franka robot, it learns from 5 minutes of play data and performs 12-step tasks from goal specifications. Code and additional analysis are available at https://symskill.github.io/ .

cs.RO↗

arXiv2026

Transformer-based prediction of two-dimensional material electronic properties under elastic strain engineering

Strain engineering provides a powerful route for tuning the electronic properties of two-dimensional (2D) materials, but exploring the full multidimensional strain space with density functional theory (DFT) is computationally prohibitive due to the nonlinear coupling between normal and shear components. In this work, we introduce a Transformer-based, multi-target surrogate model framework that achieves DFT-level bandgap prediction accuracy, reaching a mean absolute error of 0.0103 eV while retaining full interpretability through attention-weight analysis. The learned self-attention map consistently identifies shear strain as the interaction center that influences both bandgap and phonon stability, an insight not readily captured by classical feature-importance metrics. This work establishes attention-based architectures as physically interpretable surrogate models for multi-property prediction, offering a generalizable strategy for accelerating deep elastic strain engineering in materials informatics.

cond-mat.mtrl-sci↗

arXiv2026

Hierarchical Multi-Modal Planning for Fixed-Altitude Sparse Target Search and Sampling

Efficient monitoring of sparse benthic phenomena, such as coral colonies, presents a great challenge for Autonomous Underwater Vehicles. Traditional exhaustive coverage strategies are energy-inefficient, while recent adaptive sampling approaches rely on costly vertical maneuvers. To address these limitations, we propose HIMoS (Hierarchical Informative Multi-Modal Search), a fixed-altitude framework for sparse coral search-and-sample missions. The system integrates a heterogeneous sensor suite within a two-layer planning architecture. At the strategic level, a Global Planner optimizes topological routes to maximize potential discovery. At the tactical level, a receding-horizon Local Planner leverages differentiable belief propagation to generate kinematically feasible trajectories that balance acoustic substrate exploration, visual coral search, and close-range sampling. Validated in high-fidelity simulations derived from real-world coral reef benthic surveys, our approach demonstrates superior mission efficiency compared to state-of-the-art baselines.

cs.RO↗

arXiv2026

Hydrostatic Pressure-enhanced correlated magnetism and Chern insulator in moir'e WSe2

Moiré semiconductors offer flat bands where Coulomb interactions and band topology intertwine, while interlayer coupling plays a central role in forming the moiré potential. However, limited interlayer coupling strength and the lack of efficient tuning methods hinder further exploration of correlated phenomena in moiré semiconductors. Here we introduce a cryogenic dual-gated diamond-anvil platform using helium as a pressure medium, enabling reversible hydrostatic tuning together with magneto-optical spectroscopy in twisted bilayer WSe2. Pressure enhances the moiré potential, redshifts excitons, and stabilizes Stoner ferromagnetism otherwise absent at a 3.1-degree twist. Simultaneously, the half-filled C = 1 Chern insulating state strengthens, exhibiting a reduced saturation field. Moreover, we observe a topological phase transition from a Chern insulator to a Mott insulator at around 2 GPa. First-principles calculations reveal that a Gamma-to-K valence-band-maximum switching drives this transition by converting an Ising-like topological K-valley miniband into a spin-degenerate trivial Gamma miniband. Our findings demonstrate hydrostatic pressure as a powerful, continuous control axis for correlated magnetism and topological band engineering in moiré materials.

cond-mat.mtrl-sci↗

arXiv2026

Sequential Operating Simulation of Solid State Transformer-Driven Next-Generation 800 VDC Data Center

Artificial-intelligence (AI) workloads are driving rapid growth in data-center electricity use and rack power density, increasing demand for power-delivery systems that are efficient and robust to fast load transients. Conventional uninterruptible power supply (UPS) based AC distribution chains involve multiple conversion stages and line-frequency transformers, which compound losses and are less compatible with dynamic AI power profiles. Although solid-state transformers (SSTs) and 800 VDC distribution architecture are widely discussed, implementable topology/control details, and long-horizon validation with realistic operating profiles remain limited. This paper develops an SST-driven 800 VDC architecture that converts 10 kV MVAC to an 800V LVDC bus using a three-phase H-bridge AC/DC stage cascaded with a dual-active-bridge (DAB) DC/DC stage. A coordinated closed-loop control scheme, combining rectifier voltage/current regulation and DAB phase-shift control, is designed to maintain DC-bus voltage stability. The proposed system is implemented on the real-time digital simulation (RTDS) platform and evaluated via sequential simulations using real-world day- and month-scale operating profiles of data centers, benchmarked against a UPS supply chain. Numerical studies demonstrate tight 800 VDC regulation, reduced input-side energy consumption compared with the UPS baseline, and satisfactory power-quality performance. A capacitance sensitivity test quantifies tradeoffs between DC-bus ripple and low-frequency input-power oscillations, yielding a practical capacitance range for design. Overall, the work provides a reproducible evaluation workflow and actionable guidance for next-generation AI data centers.

eess.SY↗

arXiv2026

Exploring the Use of VLMs for Navigation Assistance for People with Blindness and Low Vision

This paper investigates the potential of vision-language models (VLMs) to assist people with blindness and low vision (pBLV) in navigation tasks. We evaluate state-of-the-art closed-source models, including GPT-4V, GPT-4o, Gemini-1.5-Pro, and Claude-3.5-Sonnet, alongside open-source models, such as Llava-v1.6-mistral and Llava-onevision-qwen, to analyze their capabilities in foundational visual skills: counting ambient obstacles, relative spatial reasoning, and common-sense wayfinding-pertinent scene understanding. We further assess their performance in navigation scenarios, using pBLV-specific prompts designed to simulate real-world assistance tasks. Our findings reveal notable performance disparities between these models: GPT-4o consistently outperforms others across all tasks, particularly in spatial reasoning and scene understanding. In contrast, open-source models struggle with nuanced reasoning and adaptability in complex environments. Common challenges include difficulties in accurately counting objects in cluttered settings, biases in spatial reasoning, and a tendency to prioritize object details over spatial feedback, limiting their usability for pBLV in navigation tasks. Despite these limitations, VLMs show promise for wayfinding assistance when better aligned with human feedback and equipped with improved spatial reasoning. This research provides actionable insights into the strengths and limitations of current VLMs, guiding developers on effectively integrating VLMs into assistive technologies while addressing key limitations for enhanced usability.

cs.CV↗

arXiv2025

Let Samples Speak: Mitigating Spurious Correlation by Exploiting the Clusterness of Samples

Deep learning models are known to often learn features that spuriously correlate with the class label during training but are irrelevant to the prediction task. Existing methods typically address this issue by annotating potential spurious attributes, or filtering spurious features based on some empirical assumptions (e.g., simplicity of bias). However, these methods may yield unsatisfactory performance due to the intricate and elusive nature of spurious correlations in real-world data. In this paper, we propose a data-oriented approach to mitigate the spurious correlation in deep learning models. We observe that samples that are influenced by spurious features tend to exhibit a dispersed distribution in the learned feature space. This allows us to identify the presence of spurious features. Subsequently, we obtain a bias-invariant representation by neutralizing the spurious features based on a simple grouping strategy. Then, we learn a feature transformation to eliminate the spurious features by aligning with this bias-invariant representation. Finally, we update the classifier by incorporating the learned feature transformation and obtain an unbiased model. By integrating the aforementioned identifying, neutralizing, eliminating and updating procedures, we build an effective pipeline for mitigating spurious correlation. Experiments on image and NLP debiasing benchmarks show an improvement in worst group accuracy of more than 20% compared to standard empirical risk minimization (ERM). Codes and checkpoints are available at https://github.com/davelee-uestc/nsf_debiasing .

cs.CV↗

arXiv2025

A Spatial Semantics and Continuity Perception Attention for Remote Sensing Water Body Change Detection

Remote sensing Water Body Change Detection (WBCD) aims to detect water body surface changes from bi-temporal images of the same geographic area. Recently, the scarcity of high spatial resolution datasets for WBCD restricts its application in urban and rural regions, which require more accurate positioning. Meanwhile, previous deep learning-based methods fail to comprehensively exploit the spatial semantic and structural information in deep features in the change detection networks. To resolve these concerns, we first propose a new dataset, HSRW-CD, with a spatial resolution higher than 3 meters for WBCD. Specifically, it contains a large number of image pairs, widely covering various water body types. Besides, a Spatial Semantics and Continuity Perception (SSCP) attention module is designed to fully leverage both the spatial semantics and structure of deep features in the WBCD networks, significantly improving the discrimination capability for water body. The proposed SSCP has three components: the Multi-Semantic spatial Attention (MSA), the Structural Relation-aware Global Attention (SRGA), and the Channel-wise Self-Attention (CSA). The MSA enhances the spatial semantics of water body features and provides precise spatial semantic priors for the CSA. Then, the SRGA further extracts spatial structure to learn the spatial continuity of the water body. Finally, the CSA utilizes the spatial semantic and structural priors from the MSA and SRGA to compute the similarity across channels. Specifically designed as a plug-and-play module for water body deep features, the proposed SSCP allows integration into existing WBCD models. Numerous experiments conducted on the proposed HSRW-CD and Water-CD datasets validate the effectiveness and generalization of the SSCP. The code of this work and the HSRW-CD dataset will be accessed at https://github.com/QingMa1/SSCP.

cs.CV↗

arXiv2025

Make Both Ends Meet: A Synergistic Optimization Infrared Small Target Detection with Streamlined Computational Overhead

Infrared small target detection(IRSTD) is widely recognized as a challenging task due to the inherent limitations of infrared imaging, including low signal-to-noise ratios, lack of texture details, and complex background interference. While most existing methods model IRSTD as a semantic segmentation task, but they suffer from two critical drawbacks: (1)blurred target boundaries caused by long-distance imaging dispersion; and (2) excessive computational overhead due to indiscriminate feature stackin. To address these issues, we propose the Lightweight Efficiency Infrared Small Target Detection (LE-IRSTD), a lightweight and efficient framework based on YOLOv8n, with following key innovations. Firstly, we identify that the multiple bottleneck structures within the C2f component of the YOLOv8-n backbone contribute to an increased computational burden. Therefore, we implement the Mobile Inverted Bottleneck Convolution block (MBConvblock) and Bottleneck Structure block (BSblock) in the backbone, effectively balancing the trade-off between computational efficiency and the extraction of deep semantic information. Secondly, we introduce the Attention-based Variable Convolution Stem (AVCStem) structure, substituting the final convolution with Variable Kernel Convolution (VKConv), which allows for adaptive convolutional kernels that can transform into various shapes, facilitating the receptive field for the extraction of targets. Finally, we employ Global Shuffle Convolution (GSConv) to shuffle the channel dimension features obtained from different convolutional approaches, thereby enhancing the robustness and generalization capabilities of our method. Experimental results demonstrate that our LE-IRSTD method achieves compelling results in both accuracy and lightweight performance, outperforming several state-of-the-art deep learning methods.

eess.IV↗

arXiv2025

Exploring State Space Model in Wavelet Domain: An Infrared and Visible Image Fusion Network via Wavelet Transform and State Space Model

Deep learning techniques have revolutionized the infrared and visible image fusion (IVIF), showing remarkable efficacy on complex scenarios. However, current methods do not fully combine frequency domain features with global semantic information, which will result in suboptimal extraction of global features across modalities and insufficient preservation of local texture details. To address these issues, we propose Wavelet-Mamba (W-Mamba), which integrates wavelet transform with the state-space model (SSM). Specifically, we introduce Wavelet-SSM module, which incorporates wavelet-based frequency domain feature extraction and global information extraction through SSM, thereby effectively capturing both global and local features. Additionally, we propose a cross-modal feature attention modulation, which facilitates efficient interaction and fusion between different modalities. The experimental results indicate that our method achieves both visually compelling results and superior performance compared to current state-of-the-art methods. Our code is available at https://github.com/Lmmh058/W-Mamba.

cs.CV↗

arXiv2024

LAM-YOLO: Drones-based Small Object Detection on Lighting-Occlusion Attention Mechanism YOLO

Drone-based target detection presents inherent challenges, such as the high density and overlap of targets in drone-based images, as well as the blurriness of targets under varying lighting conditions, which complicates identification. Traditional methods often struggle to recognize numerous densely packed small targets under complex background. To address these challenges, we propose LAM-YOLO, an object detection model specifically designed for drone-based. First, we introduce a light-occlusion attention mechanism to enhance the visibility of small targets under different lighting conditions. Meanwhile, we incroporate incorporate Involution modules to improve interaction among feature layers. Second, we utilize an improved SIB-IoU as the regression loss function to accelerate model convergence and enhance localization accuracy. Finally, we implement a novel detection strategy that introduces two auxiliary detection heads for identifying smaller-scale targets.Our quantitative results demonstrate that LAM-YOLO outperforms methods such as Faster R-CNN, YOLOv9, and YOLOv10 in terms of mAP@0.5 and mAP@0.5:0.95 on the VisDrone2019 public dataset. Compared to the original YOLOv8, the average precision increases by 7.1\%. Additionally, the proposed SIB-IoU loss function shows improved faster convergence speed during training and improved average precision over the traditional loss function.

cs.CV↗

arXiv2022

Optimal Rejection Function Meets Character Recognition Tasks

In this paper, we propose an optimal rejection method for rejecting ambiguous samples by a rejection function. This rejection function is trained together with a classification function under the framework of Learning-with-Rejection (LwR). The highlights of LwR are: (1) the rejection strategy is not heuristic but has a strong background from a machine learning theory, and (2) the rejection function can be trained on an arbitrary feature space which is different from the feature space for classification. The latter suggests we can choose a feature space that is more suitable for rejection. Although the past research on LwR focused only on its theoretical aspect, we propose to utilize LwR for practical pattern classification tasks. Moreover, we propose to use features from different CNN layers for classification and rejection. Our extensive experiments of notMNIST classification and character/non-character classification demonstrate that the proposed method achieves better performance than traditional rejection strategies.

cs.CV↗

arXiv2021

Athermal Activation in Glassy Fluid

In this article, the mechanism of the unexpected high fluidity in SiOx nanowire under modest irradiation was proposed, the high fluidity is attributed to the long lifetime of irradiation-induced holes, which arise from formation of small polarons. The holes created in irradiation could have a long lifetime, and localized in space, such missing of bonding electron could suppress the energy barrier(athermal activation effect) for a Pachner move of the network. The atomic level dynamics of the system is proposed by interaction of phonon and local configuration, the activation effect was then studied with passing rate of corresponding stochastic dynamic equation, calculation shows an exponential dependent of the time-lapse of Pachner move to lifetime of the activation, furthermore, connection between the local configuration time and viscosity of the fluid indicates a strong sensitivity of viscosity to lifetime of the athermal activation, such mechanism would give an effective interpretation to the unexpected high fluidity together with the passivation effect of the conductor on the material.

cond-mat.soft↗

arXiv2020

Regularized Pooling

In convolutional neural networks (CNNs), pooling operations play important roles such as dimensionality reduction and deformation compensation. In general, max pooling, which is the most widely used operation for local pooling, is performed independently for each kernel. However, the deformation may be spatially smooth over the neighboring kernels. This means that max pooling is too flexible to compensate for actual deformations. In other words, its excessive flexibility risks canceling the essential spatial differences between classes. In this paper, we propose regularized pooling, which enables the value selection direction in the pooling operation to be spatially smooth across adjacent kernels so as to compensate only for actual deformations. The results of experiments on handwritten character images and texture images showed that regularized pooling not only improves recognition accuracy but also accelerates the convergence of learning compared with conventional pooling operations.

cs.CV↗

arXiv2017

Variable Partitioning for Distributed Optimization

This paper is about how to partition decision variables while decomposing a large-scale optimization problem for the best performance of distributed solution methods. Solving a large-scale optimization problem sequen- tially can be computationally challenging. One classic approach is to decompose the problem into smaller sub-problems and solve them in a distributed fashion. However, there is little discussion in the literature on which variables should be grouped together to form the sub-problems, especially when the optimization formulation involves complex constraints. We focus on one of the most popular distributed approaches, dual decomposition and distributed sub-gradient methods. Based on a theoretical guarantee on its convergence rate, we explain that a partition of variables can critically affect the speed of convergence and highlight the importance of the number of dualized constraints. Then, we introduce a novel approach to find a partition that reduces the number of dualized constraints by utilizing a community detection algorithm from physics literature. Roughly speaking, the proposed method groups decision variables that appear together in con- straints and solves the resulting sub-problems with blocks of variables in parallel. Empirical experiments on a real application show that the proposed method significantly accelerates the convergence of the distributed sub-gradient method. The advantage of our approach becomes more significant as the size of the problem increases and each constraint involves more variables.

math.OC↗

arXiv2017

On Designing of a Low Leakage Patient-Centric Provider Network

When a patient in a provider network seeks services outside of their community, the community experiences a leakage. Leakage is undesirable as it typically leads to higher out-of-network cost for patient and increases barrier for care coordination, which is particularly problematic for Accountable Care Organization (ACO) as the in-network providers are financially responsible for patient quality and outcome. We aim to design a data-driven method to identify naturally occurring provider networks driven by diabetic patient choices, and understand the relationship among provider composition, patient composition, and service leakage pattern. We construct a healthcare provider network based on patients' historical medical insurance claims. A community detection algorithm is used to identify naturally occurring communities of collaborating providers. Finally, import-export analysis is conducted to benchmark their leakage pattern and identify further leakage reduction opportunity. The design yields six major provider communities with diverse profiles. Some communities are geographically concentrated, while others tend to draw patients with certain diabetic co-morbidities. Providers from the same healthcare institution are likely to be assigned to the same community. While most communities have high within-community utilization and spending, at 85% and 86% respectively, leakage still persists. Hence, we utilize a metric from import-export analysis to detect leakage, gaining insight on how to minimizing leakage. In conclusion, we identify patient-driven provider organization by surfacing providers who share a large number of patients. By analyzing the import-export behavior of each identified community using a novel approach and profiling community patient and provider composition we understand the key features of having a balanced number of PCP and specialists and provider heterogeneity.

stat.AP↗