arXiv subjects

Yi Chen

Publications and source records attributed to Yi Chen.

At least 109 records · Page 6Linked to original sources

arXiv2026

MR-Align: Meta-Reasoning Informed Factuality Alignment for Large Reasoning Models

Large reasoning models (LRMs) show strong capabilities in complex reasoning, yet their marginal gains on evidence-dependent factual questions are limited. We find this limitation is partially attributable to a reasoning-answer hit gap, where the model identifies the correct facts during reasoning but fails to incorporate them into the final response, thereby reducing factual fidelity. To address this issue, we propose MR-ALIGN, a Meta-Reasoning informed alignment framework that enhances factuality without relying on external verifiers. MR-ALIGN quantifies state transition probabilities along the model's thinking process and constructs a transition-aware implicit reward that reinforces beneficial reasoning patterns while suppressing defective ones at the atomic thinking segments. This re-weighting reshapes token-level signals into probability-aware segment scores, encouraging coherent reasoning trajectories that are more conducive to factual correctness. Empirical evaluations across four factual QA datasets and one long-form factuality benchmark show that MR-ALIGN consistently improves accuracy and truthfulness while reducing misleading reasoning. These results highlight that aligning the reasoning process itself, rather than merely the outputs, is pivotal for advancing factuality in LRMs.

cs.CL↗

arXiv2026

Inconsistencies in Classification of Online News Articles: A Call for Common Standards in Brand Safety Services

This study examines inconsistencies in the brand safety classifications of online news articles by analyzing ratings from three leading brand safety providers, DoubleVerify, Integral Ad Science, and Oracle. We focus on news content because of its central role in public discourse and the significant financial consequences of unsafe classifications in a sector that is already underserved by digital ad spending. By collecting data from 4,352 news articles on 51 domains, our analysis shows that brand safety services often produce conflicting classifications, with significant discrepancies between providers. These inconsistencies can have harmful consequences for both advertisers and publishers, leading to misplaced advertising spending and revenue losses. This research provides critical insights into the shortcomings of the current brand safety landscape. We argue for a standardized and transparent brand safety system to mitigate the harmful effects of the current system on the digital advertising ecosystem.

cs.CY↗

arXiv2025

Evidence of Mott Insulator with Thermally Induced Melting Behavior in Kagome Compound Nb3Cl8

The kagome lattice provides a playground to explore novel correlated quantum states due to the presence of flat bands in its electronic structure. Recently discovered layered kagome compound Nb3Cl8 has been proposed as a Mott insulator coming from the half-filled flat band. Here we have carried out systematic transport study to uncover the evidence of Mott insulator in Nb3Cl8 thin flakes. Bipolar semiconducting property with Fermi level close to conduction band has been revealed. We have further probed the chemical potential of Nb3Cl8 by tracing the charge neutrality point of the monolayer graphene proximate to Nb3Cl8. The gap of Nb3Cl8 flakes is approximately 1.10 eV at 100 K and shows pronounced temperature dependence, decreasing substantially with increasing temperature to ~0.63 eV at 300 K. The melting behavior of the gapped state is in consistent with theoretically proposed Mott insulator in Nb3Cl8. Our work has demonstrated Nb3Cl8 as a promising platform to study strongly correlated physics at relatively high temperature.

cond-mat.str-el↗

arXiv2025

CD-DPE: Dual-Prompt Expert Network Based on Convolutional Dictionary Feature Decoupling for Multi-Contrast MRI Super-Resolution

Multi-contrast magnetic resonance imaging (MRI) super-resolution intends to reconstruct high-resolution (HR) images from low-resolution (LR) scans by leveraging structural information present in HR reference images acquired with different contrasts. This technique enhances anatomical detail and soft tissue differentiation, which is vital for early diagnosis and clinical decision-making. However, inherent contrasts disparities between modalities pose fundamental challenges in effectively utilizing reference image textures to guide target image reconstruction, often resulting in suboptimal feature integration. To address this issue, we propose a dual-prompt expert network based on a convolutional dictionary feature decoupling (CD-DPE) strategy for multi-contrast MRI super-resolution. Specifically, we introduce an iterative convolutional dictionary feature decoupling module (CD-FDM) to separate features into cross-contrast and intra-contrast components, thereby reducing redundancy and interference. To fully integrate these features, a novel dual-prompt feature fusion expert module (DP-FFEM) is proposed. This module uses a frequency prompt to guide the selection of relevant reference features for incorporation into the target image, while an adaptive routing prompt determines the optimal method for fusing reference and target features to enhance reconstruction quality. Extensive experiments on public multi-contrast MRI datasets demonstrate that CD-DPE outperforms state-of-the-art methods in reconstructing fine details. Additionally, experiments on unseen datasets demonstrated that CD-DPE exhibits strong generalization capabilities.

cs.CV↗

arXiv2025

ChartAgent: A Chart Understanding Framework with Tool Integrated Reasoning

With their high information density and intuitive readability, charts have become the de facto medium for data analysis and communication across disciplines. Recent multimodal large language models (MLLMs) have made notable progress in automated chart understanding, yet they remain heavily dependent on explicit textual annotations and the performance degrades markedly when key numerals are absent. To address this limitation, we introduce ChartAgent, a chart understanding framework grounded in Tool-Integrated Reasoning (TIR). Inspired by human cognition, ChartAgent decomposes complex chart analysis into a sequence of observable, replayable steps. Supporting this architecture is an extensible, modular tool library comprising more than a dozen core tools, such as keyelement detection, instance segmentation, and optical character recognition (OCR), which the agent dynamically orchestrates to achieve systematic visual parsing across diverse chart types. Leveraging TIRs transparency and verifiability, ChartAgent moves beyond the black box paradigm by standardizing and consolidating intermediate outputs into a structured Evidence Package, providing traceable and reproducible support for final conclusions. Experiments show that ChartAgent substantially improves robustness under sparse annotation settings, offering a practical path toward trustworthy and extensible systems for chart understanding.

cs.CV↗

arXiv2025

Dora: QoE-Aware Hybrid Parallelism for Distributed Edge AI

With the proliferation of edge AI applications, satisfying user quality of experience (QoE) requirements, such as model inference latency, has become a first class objective, as these models operate in resource constrained settings and directly interact with users. Yet, modern AI models routinely exceed the resource capacity of individual devices, necessitating distributed execution across heterogeneous devices over variable and contention prone networks. Existing planners for hybrid (e.g., data and pipeline) parallelism largely optimize for throughput or device utilization, overlooking QoE, leading to severe resource inefficiency (e.g., unnecessary energy drain) or QoE violations under runtime dynamics. We present Dora, a framework for QoE aware hybrid parallelism in distributed edge AI training and inference. Dora jointly optimizes heterogeneous computation, contention prone networks, and multi dimensional QoE objectives via three key mechanisms: (i) a heterogeneity aware model partitioner that determines and assigns model partitions across devices, forming a compact set of QoE compliant plans; (ii) a contention aware network scheduler that further refines these candidate plans by maximizing compute communication overlap; and (iii) a runtime adapter that adaptively composes multiple plans to maximize global efficiency while respecting overall QoEs. Across representative edge deployments, including smart homes, traffic analytics, and small edge clusters, Dora achieves 1.1--6.3 times faster execution and, alternatively, reduces energy consumption by 21--82 percent, all while maintaining QoE under runtime dynamics.

cs.DC↗

arXiv2025

Rhea: Role-aware Heuristic Episodic Attention for Conversational LLMs

Large Language Models (LLMs) have achieved remarkable performance on single-turn tasks, yet their effectiveness deteriorates in multi-turn conversations. We define this phenomenon as cumulative contextual decay - a progressive degradation of contextual integrity caused by attention pollution, dilution, and drift. To address this challenge, we propose Rhea (Role-aware Heuristic Episodic Attention), a novel framework that decouples conversation history into two functionally independent memory modules: (1) an Instructional Memory (IM) that persistently stores high-fidelity global constraints via a structural priority mechanism, and (2) an Episodic Memory (EM) that dynamically manages user-model interactions via asymmetric noise control and heuristic context retrieval. During inference, Rhea constructs a high signal-to-noise context by applying its priority attention: selectively integrating relevant episodic information while always prioritizing global instructions. To validate this approach, experiments on multiple multi-turn conversation benchmarks - including MT-Eval and Long-MT-Bench+ - show that Rhea mitigates performance decay and improves overall accuracy by 1.04 points on a 10-point scale (a 16% relative gain over strong baselines). Moreover, Rhea maintains near-perfect instruction fidelity (IAR > 8.1) across long-horizon interactions. These results demonstrate that Rhea provides a principled and effective framework for building more precise, instruction-consistent conversational LLMs.

cs.CL↗

arXiv2025

Discover, Learn, and Reinforce: Scaling Vision-Language-Action Pretraining with Diverse RL-Generated Trajectories

Scaling vision-language-action (VLA) model pre-training requires large volumes of diverse, high-quality manipulation trajectories. Most current data is obtained via human teleoperation, which is expensive and difficult to scale. Reinforcement learning (RL) methods learn useful skills through autonomous exploration, making them a viable approach for generating data. However, standard RL training collapses to a narrow execution pattern, limiting its utility for large-scale pre-training. We propose Discover, Lea rn and Reinforce (DLR), an information-theoretic pattern discovery framework that generates multiple distinct, high-success behavioral patterns for VLA pretraining. Empirically, DLR generates a markedly more diverse trajectory corpus on LIBERO. Specifically, it learns multiple distinct, high-success strategies for the same task where standard RL discovers only one, and hence it covers substantially broader regions of the state-action space. When adapted to unseen downstream task suites, VLA models pretrained on our diverse RL data surpass counterparts trained on equal-sized standard RL datasets. Moreover, DLR exhibits positive data-scaling behavior that single-pattern RL lacks. These results position multi-pattern RL as a practical, scalable data engine for embodied foundation models.

cs.RO↗

arXiv2025

MeteorPred: A Meteorological Multimodal Large Model and Dataset for Severe Weather Event Prediction

Timely and accurate forecasts of severe weather events are essential for early warning and for constraining downstream analysis and decision-making. Since severe weather events prediction still depends on subjective, time-consuming expert interpretation, end-to-end "AI weather station" systems are emerging but face three major challenges: (1) scarcity of severe weather event samples; (2) imperfect alignment between high-dimensional meteorological data and textual warnings; (3) current multimodal language models cannot effectively process high-dimensional meteorological inputs or capture their complex spatiotemporal dependencies. To address these challenges, we introduce MP-Bench, the first large-scale multimodal dataset for severe weather events prediction, comprising 421,363 pairs of raw multi-year meteorological data and corresponding text caption, covering a wide range of severe weather scenarios. On top of this dataset, we develop a Meteorology Multimodal Large Model (MMLM) that directly ingests 4D meteorological inputs. In addition, it is designed to accommodate the unique characteristics of 4D meteorological data flow, incorporating three plug-and-play adaptive fusion modules that enable dynamic feature extraction and integration across temporal sequences, vertical pressure layers, and spatial dimensions. Extensive experiments on MP-Bench show that MMLM achieves strong performance across multiple tasks, demonstrating effective severe weather understanding and representing a key step toward automated, AI-driven severe weather events forecasting systems. Our source code and dataset will be made publicly available.

cs.AI↗

arXiv2025

TOPP-DWR: Time-Optimal Path Parameterization of Differential-Driven Wheeled Robots Considering Piecewise-Constant Angular Velocity Constraints

Differential-driven wheeled robots (DWR) represent the quintessential type of mobile robots and find extensive appli- cations across the robotic field. Most high-performance control approaches for DWR explicitly utilize the linear and angular velocities of the trajectory as control references. However, existing research on time-optimal path parameterization (TOPP) for mobile robots usually neglects the angular velocity and joint vel- ocity constraints, which can result in degraded control perfor- mance in practical applications. In this article, a systematic and practical TOPP algorithm named TOPP-DWR is proposed for DWR and other mobile robots. First, the non-uniform B-spline is adopted to represent the initial trajectory in the task space. Second, the piecewise-constant angular velocity, as well as joint velocity, linear velocity, and linear acceleration constraints, are incorporated into the TOPP problem. During the construction of the optimization problem, the aforementioned constraints are uniformly represented as linear velocity constraints. To boost the numerical computational efficiency, we introduce a slack variable to reformulate the problem into second-order-cone programming (SOCP). Subsequently, comparative experiments are conducted to validate the superiority of the proposed method. Quantitative performance indexes show that TOPP-DWR achieves TOPP while adhering to all constraints. Finally, field autonomous navigation experiments are carried out to validate the practicability of TOPP-DWR in real-world applications.

cs.RO↗

arXiv2025

Tunable Luttinger liquid and correlated insulating states in one-dimensional moiré superlattices

Two-dimensional moiré superlattices have been extensively studied, and a variety of correlated phenomena have been observed. However, their lower-dimensional counterpart, one-dimensional (1D) moiré superlattices, remain largely unexplored. Electrons in 1D are generally described by Luttinger liquid theory, with universal scaling relations depending only on the Luttinger parameter g. In particular, at half-filling, Umklapp scattering plays a crucial role, as it can significantly change the conductance-temperature scaling relation and lead to Mott insulators. However, this prediction has never been observed since doping an empty band to half-filling was extremely difficult. Here, we show that the marriage of moiré superlattices and 1D electrons makes it possible to study the Luttinger liquid in an exceptionally wide filling region simply by electrical gating. We perform transport measurements on 1D moiré superlattices of carbon nanotubes on hexagonal boron nitride (hBN) substrates, and observe correlated insulating states at 1/4 and 1/2 fillings of the superlattice mini-band, where Umklapp scattering becomes dominant. We also observe a T-linear conductance at these commensurate fillings over a range of temperatures. Strikingly, the T-linear conductance leads to a strongly suppressed Luttinger parameter, suggesting a state of extreme correlation.

cond-mat.str-el↗

arXiv2025

How Do VLAs Effectively Inherit from VLMs?

Vision-language-action (VLA) models hold the promise to attain generalizable embodied control. To achieve this, a pervasive paradigm is to leverage the rich vision-semantic priors of large vision-language models (VLMs). However, the fundamental question persists: How do VLAs effectively inherit the prior knowledge from VLMs? To address this critical question, we introduce a diagnostic benchmark, GrinningFace, an emoji tabletop manipulation task where the robot arm is asked to place objects onto printed emojis corresponding to language instructions. This task design is particularly revealing -- knowledge associated with emojis is ubiquitous in Internet-scale datasets used for VLM pre-training, yet emojis themselves are largely absent from standard robotics datasets. Consequently, they provide a clean proxy: successful task completion indicates effective transfer of VLM priors to embodied control. We implement this diagnostic task in both simulated environment and a real robot, and compare various promising techniques for knowledge transfer. Specifically, we investigate the effects of parameter-efficient fine-tuning, VLM freezing, co-training, predicting discretized actions, and predicting latent actions. Through systematic evaluation, our work not only demonstrates the critical importance of preserving VLM priors for the generalization of VLA but also establishes guidelines for future research in developing truly generalizable embodied AI systems.

cs.RO↗

arXiv2025

Energy Correlators from Partons to Hadrons: Unveiling the Dynamics of the Strong Interactions with Archival ALEPH Data

Quantum Chromodynamics (QCD) is a remarkably rich theory exhibiting numerous emergent degrees of freedom, from flux tubes to hadrons. Their description in terms of the underlying quarks and gluons of the QCD Lagrangian remains a central challenge of modern physics. Colliders offer a unique opportunity to probe these phenomena experimentally: high energy partons produced from the QCD vacuum excite these emergent degrees, imprinting their dynamics in correlations in asymptotic energy flux. Decoding these correlations requires measurements with exceptional angular resolution, beyond that achieved in previous measurements. Recent progress has enabled precision calculations of energy flux on charged particles alone, allowing data-theory comparisons for measurements using high resolution tracking detectors. In this Letter, we resurrect thirty-year-old data from the ALEPH tracker, and perform a high angular resolution measurement of the two-point correlation of energy flux, probing QCD over three orders of magnitude in scale in a single measurement. Our measurement unveils for the first time the full spectrum of the correlator, including light-ray quasi-particle states, flux-tube excitations, and their transitions into confined hadrons. We compare our measurement with record precision theoretical predictions, achieving percent level agreement, and revealing interesting new phenomena in the confinement transitions. More broadly, we highlight the immense potential of this newly unlocked archival data set, the so called "recycling frontier", and emphasize synergies with ongoing and future collider experiments.

hep-ph↗

arXiv2025

Using Salient Object Detection to Identify Manipulative Cookie Banners that Circumvent GDPR

The main goal of this paper is to study how often cookie banners that comply with the General Data Protection Regulation (GDPR) contain aesthetic manipulation, a design tactic to draw users' attention to the button that permits personal data sharing. As a byproduct of this goal, we also evaluate how frequently the banners comply with GDPR and the recommendations of national data protection authorities regarding banner designs. We visited 2,579 websites and identified the type of cookie banner implemented. Although 45% of the relevant websites have fully compliant banners, we found aesthetic manipulation on 38% of the compliant banners. Unlike prior studies of aesthetic manipulation, we use a computer vision model for salient object detection to measure how salient (i.e., attention-drawing) each banner element is. This enables the discovery of new types of aesthetic manipulation (e.g., button placement), and leads us to conclude that aesthetic manipulation is more common than previously reported (38% vs 27% of banners). To study the effects of user and/or website location on cookie banner design, we include websites within the European Union (EU), where privacy regulation enforcement is more stringent, and websites outside the EU. We visited websites from IP addresses in the EU and from IP addresses in the United States (US). We find that 13.9% of EU websites change their banner design when the user is from the US, and EU websites are roughly 48.3% more likely to use aesthetic manipulation than non-EU websites, highlighting their innovative responses to privacy regulation.

cs.CY↗

arXiv2025

A Study on Privacy-Preserving Scholarship Evaluation Based on Decentralized Identity and Zero-Knowledge Proofs

Traditional centralized scholarship evaluation processes typically require students to submit detailed academic records and qualification information, which exposes them to risks of data leakage and misuse, making it difficult to simultaneously ensure privacy protection and transparent auditability. To address these challenges, this paper proposes a scholarship evaluation system based on Decentralized Identity (DID) and Zero-Knowledge Proofs (ZKP). The system aggregates multidimensional ZKPs off-chain, and smart contracts verify compliance with evaluation criteria without revealing raw scores or computational details. Experimental results demonstrate that the proposed solution not only automates the evaluation efficiently but also maximally preserves student privacy and data integrity, offering a practical and trustworthy technical paradigm for higher education scholarship programs.

cs.CR↗

arXiv2025

DeshadowMamba: Deshadowing as 1D Sequential Similarity

Recent deep models for image shadow removal often rely on attention-based architectures to capture long-range dependencies. However, their fixed attention patterns tend to mix illumination cues from irrelevant regions, leading to distorted structures and inconsistent colors. In this work, we revisit shadow removal from a sequence modeling perspective and explore the use of Mamba, a selective state space model that propagates global context through directional state transitions. These transitions yield an efficient global receptive field while preserving positional continuity. Despite its potential, directly applying Mamba to image data is suboptimal, since it lacks awareness of shadow-non-shadow semantics and remains susceptible to color interference from nearby regions. To address these limitations, we propose CrossGate, a directional modulation mechanism that injects shadow-aware similarity into Mamba's input gate, allowing selective integration of relevant context along transition axes. To further ensure appearance fidelity, we introduce ColorShift regularization, a contrastive learning objective driven by global color statistics. By synthesizing structured informative negatives, it guides the model to suppress color contamination and achieve robust color restoration. Together, these components adapt sequence modeling to the structural integrity and chromatic consistency required for shadow removal. Extensive experiments on public benchmarks demonstrate that DeshadowMamba achieves state-of-the-art visual quality and strong quantitative performance.

cs.CV↗

arXiv2025

Unbinned measurement of thrust in $e^+e^-$ collisions at $\sqrt{s}$ = 91.2 GeV with ALEPH archived data

The strong coupling constant ($α_{S}$) is a fundamental parameter of quantum chromodynamics (QCD), the theory of the strong force. Some of the earliest precise constraints on $α_{S}$ came from measurements of event shape observables, such as thrust ($T$), using hadronic $Z$ boson decays produced in $e^+e^-$ collisions. However, recent work has revealed discrepancies between event-shape-based extractions of $α_{S}$ and values determined using other experimental methods. This work reexamines archived $e^+e^-$ data collected at a collision energy of $\sqrt{s}=91.2$ GeV by the ALEPH detector at the Large Electron-Positron Collider. Modern machine learning techniques are used to correct for detector effects in an unbinned manner, allowing the $T$ distribution to be measured with higher granularity than previous ALEPH measurements. The new measurement reveals a small but systematic shift towards larger values of $τ=1-T$, and the potential implications of this shift for $α_{S}$ extractions are illustrated by comparing to state-of-the-art theoretical calculations. In addition, the region of $-6<\logτ<-2$, where poorly-understood non-perturbative effects are large, is compared to modern parton shower Monte Carlo simulations. This measurement provides unique new inputs for $α_{S}$ extractions and also improves constraints on phenomenological models of QCD dynamics such as parton fragmentation and hadronization.

hep-ex↗

arXiv2025

Graph Unlearning Meets Influence-aware Negative Preference Optimization

Recent advancements in graph unlearning models have enhanced model utility by preserving the node representation essentially invariant, while using gradient ascent on the forget set to achieve unlearning. However, this approach causes a drastic degradation in model utility during the unlearning process due to the rapid divergence speed of gradient ascent. In this paper, we introduce \textbf{INPO}, an \textbf{I}nfluence-aware \textbf{N}egative \textbf{P}reference \textbf{O}ptimization framework that focuses on slowing the divergence speed and improving the robustness of the model utility to the unlearning process. Specifically, we first analyze that NPO has slower divergence speed and theoretically propose that unlearning high-influence edges can reduce impact of unlearning. We design an influence-aware message function to amplify the influence of unlearned edges and mitigate the tight topological coupling between the forget set and the retain set. The influence of each edge is quickly estimated by a removal-based method. Additionally, we propose a topological entropy loss from the perspective of topology to avoid excessive information loss in the local structure during unlearning. Extensive experiments conducted on five real-world datasets demonstrate that INPO-based model achieves state-of-the-art performance on all forget quality metrics while maintaining the model's utility. Codes are available at \href{https://github.com/sh-qiangchen/INPO}{https://github.com/sh-qiangchen/INPO}.

cs.LG↗