arXiv subjects

Rui Dong

Publications and source records attributed to Rui Dong.

At least 19 recordsLinked to original sources

arXiv2026

When Language Models Meet NeuroGraphs: Exploring Enhanced Agentic LLM Framework Towards Brain Network Analysis

Brain network analysis is crucial for understanding cognition and neurological disorders, yet existing deep learning methods mainly treat connectome analysis as a graph-to-logit classification problem, offering limited explanatory reasoning. Large language models (LLMs) provide a promising interface for knowledge-intensive scientific analysis, but directly applying general-purpose LLMs to brain networks remains challenging due to the structure-language gap, limited neuroscience grounding, and overconfident positive predictions. In this paper, we propose \textbf{BrainAgent}, an agentic LLM framework for knowledge-enhanced brain network analysis. BrainAgent reformulates connectome classification as an iterative process of topology-aware understanding, external retrieval, reasoning, and reflection. Specifically, it first converts raw brain networks into compact multi-level structural descriptions through brain-specific analysis tools, then retrieves relevant neuroscience knowledge and task-specific cases to ground the reasoning process, and finally generates structured predictions with reflective verification. Experiments on four public rs-fMRI datasets show that BrainAgent consistently improves different closed-source and open-source LLM backbones over direct prompting and standard reasoning baselines. Further ablation and interpretability analyses demonstrate the effectiveness of each component and show that BrainAgent produces more comprehensive, multi-level, and verifiable explanations.These results indicate that agentic LLMs provide a practical route toward interpretable and knowledge-grounded brain network analysis.

cs.MA↗

arXiv2026

Single-Contact Problem in Atomically Flat Interfaces: a Simulation Approach

Understanding friction at single-asperity contacts is essential for bridging the gap between nanoscale structural superlubricity and realistic tribological systems dominated by Hertzian contact geometry. In this work, we combine atomistic simulations and a modified continuum model to investigate the onset of sliding at crystalline SiO$_2$/SiO$_2$ interfaces. Interfacial sliding potential energy surfaces (ISPES) are computed to determine the load-dependent shear strength and minimal-scale sliding (MSS) friction. Both quantities exhibit linear dependence on normal pressure below 3 GPa, and have non-zero values at zero pressure. Incorporating these parameters, we extend the classical Mindlin model by including adhesion and nanoscale load effects, allowing us to describe the stick to slip transition under realistic Hertzian stress distributions. The model shows that nonuniform pressure distributions substantially lower the effective static friction, and oscillatory-shear experiments on graphene-passivated contacts reproduce both the predicted stiffness-collapse signature and, in the passivated limit, the adhesion-limited shear strength obtained from simulation, supporting the model's relevance to real micro-asperity tribology.

cond-mat.mes-hall↗

arXiv2026

Cheeger Inequalities for the Persistent Laplacian

We study Cheeger-type inequalities for persistent Laplacians associated with inclusions of simplicial complexes $\mathcal{K}\hookrightarrow \mathcal{L}$. We introduce a persistent up $p$-Laplacian $\Delta_{q,p,\mathrm{up}}^{\mathcal{K},\mathcal{L}}$ for $p\geq 1$. For $p=2$, this recovers the usual persistent up Laplacian, while for $p=1$ it yields a nonzero persistent Cheeger constant $\varphi_q^{\mathcal{K},\mathcal{L}}$. We prove a Cheeger-type inequality relating $\varphi_q^{\mathcal{K},\mathcal{L}}$ to the smallest nonzero eigenvalue of $\Delta_{q,\mathrm{up}}^{\mathcal{K},\mathcal{L}}$. This gives a persistent extension of recent work by Jost and Zhang (arXiv:2302.01069). We then study two more structured settings. Under a locally complete $q$-skeleton assumption on $\mathcal{K}$, we extend the complete-skeleton isoperimetric inequality of Parzanchevski--Rosenthal--Tessler (arXiv:1207.0638) to the persistent setting. For orientable $(q+1)$-dimensional pseudomanifolds, we prove a Kron-type reduction of the persistent up Laplacian to a vertex- and edge-weighted graph Laplacian, possibly with Dirichlet boundary terms, and obtain two-sided Cheeger inequalities; this is related to the dual-graph perspective in the work of Steenbergen--Klivans--Mukherjee (arXiv:1209.5091). We also describe the nonzero persistent Cheeger constant $\varphi_q^{\mathcal{K},\mathcal{L}}$ explicitly in terms of the dual graph in the non-branching pseudomanifold case. Finally, we specialize our two constructions to graph inclusions and compare them with the graph-pair theory of M\'emoli--Wan--Wang (arXiv:2012.02808). We establish precise relationships between the two persistent Cheeger constants arising from our constructions and the corresponding constants derived via Kron reduction.

math.AT↗

arXiv2026

Presynthesis: Towards Scaling Up Program Synthesis with Finer-Grained Abstract Semantics

Abstract semantics has proven to be instrumental for accelerating search-based program synthesis, by enabling the sound pruning of a set of incorrect programs (without enumerating them). One may expect faster synthesis with increasingly finer-grained abstract semantics. Unfortunately, to the best of our knowledge, this is not the case, yet. The reason is because, as abstraction granularity increases -- while fewer programs are enumerated -- pruning becomes more costly. This imposes a fundamental limit on the overall synthesis performance, which we aim to address in this work. Our key idea is to introduce an offline presynthesis phase, which consists of two steps. Given a DSL with abstract semantics, the first semantics modeling step constructs a tree automaton A for a space of inputs -- such that, for any program P and for any considered input I, A has a run that corresponds to P's execution on I under abstract semantics. Then, the second step builds an oracle O for A. This O enables fast pruning during synthesis, by allowing us to efficiently find exactly those DSL programs that satisfy a given input-output example under abstract semantics. We have implemented this presynthesis-based synthesis paradigm in a framework, Foresighter. On top of it, we have developed three instantiations for SQL, string transformation, and matrix manipulation. All of them significantly outperform prior work in the respective domains.

cs.PL↗

arXiv2026

M3D-BFS: a Multi-stage Dynamic Fusion Strategy for Sample-Adaptive Multi-Modal Brain Network Analysis

Multi-modal fusion is of great significance in neuroscience which integrates information from different modalities and can achieve better performance than uni-modal methods in downstream tasks. Current multi-modal fusion methods in brain networks, which mainly focus on structural connectivity (SC) and functional connectivity (FC) modalities, are static in nature. They feed different samples into the same model with identical computation, ignoring inherent difference between input samples. This lack of sample adaptation hinders model's further performance. To this end, we innovatively propose a multi-stage dynamic fusion strategy (M3D-BFS) for sample-adaptive multi-modal brain network analysis. Unlike other static fusion methods, we design different mixture-of-experts (MoEs) for uni- and multi-modal representations where modules can adaptively change as input sample changes during inference. To alleviate issue of MoE where training of experts may be collapsed, we divide our method into 3 stages. We first train uni-modal encoders respectively, then pretrain single experts of MoEs before finally finetuning the whole model. A multi-modal disentanglement loss is designed to enhance the final representations. To the best of our knowledge, this is the first work for dynamic fusion for multi-modal brain network analysis. Extensive experiments on different real-world datasets demonstrates the superiority of M3D-BFS.

cs.AI↗

arXiv2026

BLEG: LLM Functions as Powerful fMRI Graph-Enhancer for Brain Network Analysis

Graph Neural Networks (GNNs) have been widely used in diverse brain network analysis tasks based on preprocessed functional magnetic resonance imaging (fMRI) data. However, their performances are constrained due to high feature sparsity and inherent limitations of domain knowledge within uni-modal neurographs. Meanwhile, large language models (LLMs) have demonstrated powerful representation capabilities. Combining LLMs with GNNs presents a promising direction for brain network analysis. While LLMs and MLLMs have emerged in neuroscience, integration of LLMs with graph-based data remains unexplored. In this work, we deal with these issues by incorporating LLM's powerful representation and generalization capabilities. Considering great cost for directly tuning LLMs, we instead function LLM as enhancer to boost GNN's performance on downstream tasks. Our method, namely BLEG, can be divided into three stages. We firstly prompt LLM to get augmented texts for fMRI graph data, then we design a LLM-LM instruction tuning method to get enhanced textual representations at a relatively lower cost. GNN is trained together for coarsened alignment. Finally we finetune an adapter after GNN for given downstream tasks. Alignment loss between LM and GNN logits is designed to further enhance GNN's representation. Extensive experiments on different datasets confirmed BLEG's superiority.Code can be available at https://github.com/KamonRiderDR/BLEG.

cs.LG↗

arXiv2025

SHERLOCK: Towards Dynamic Knowledge Adaptation in LLM-enhanced E-commerce Risk Management

Effective e-commerce risk management requires in-depth case investigations to identify emerging fraud patterns in highly adversarial environments. However, manual investigation typically requires analyzing the associations and couplings among multi-source heterogeneous data, a labor-intensive process that limits efficiency. While Large Language Models (LLMs) show promise in automating these analyses, their deployment is hindered by the complexity of risk scenarios and the sparsity of long-tail domain knowledge. To address these challenges, we propose Sherlock, a framework that integrates structured domain knowledge with LLM-based reasoning through three core modules. First, we construct a domain Knowledge Base (KB) by distilling structured expertise from heterogeneous knowledge sources. Second, we design a two-stage retrieval-augmented generation strategy tailored for case investigation, which combines input contextual augmentation with a Reflect & Refine module to fully leverage the KB for improved analysis quality. Finally, we develop an integrated platform for operations and annotation to drive a self-evolving data flywheel. By combining real-time hotfixes through KB updates with periodic logic alignment via post-training, we facilitate continuous system evolution to counteract adversarial drifts. Online A/B tests at JD dot com demonstrate that Sherlock achieves an 82% Expert Acceptance Rate (EAR) and a 386.7% increase in daily investigation throughput. An additional 90-day evaluation shows that the flywheel successfully recovers from performance decay caused by changing tactics twice, raising the EAR ceiling by around 3.5% through autonomous model updates.

cs.IR↗

arXiv2025

DITING: A Multi-Agent Evaluation Framework for Benchmarking Web Novel Translation

Large language models (LLMs) have substantially advanced machine translation (MT), yet their effectiveness in translating web novels remains unclear. Existing benchmarks rely on surface-level metrics that fail to capture the distinctive traits of this genre. To address these gaps, we introduce DITING, the first comprehensive evaluation framework for web novel translation, assessing narrative and cultural fidelity across six dimensions: idiom translation, lexical ambiguity, terminology localization, tense consistency, zero-pronoun resolution, and cultural safety, supported by over 18K expert-annotated Chinese-English sentence pairs. We further propose AgentEval, a reasoning-driven multi-agent evaluation framework that simulates expert deliberation to assess translation quality beyond lexical overlap, achieving the highest correlation with human judgments among seven tested automatic metrics. To enable metric comparison, we develop MetricAlign, a meta-evaluation dataset of 300 sentence pairs annotated with error labels and scalar quality scores. Comprehensive evaluation of fourteen open, closed, and commercial models reveals that Chinese-trained LLMs surpass larger foreign counterparts, and that DeepSeek-V3 delivers the most faithful and stylistically coherent translations. Our work establishes a new paradigm for exploring LLM-based web novel translation and provides public resources to advance future research.

cs.CL↗

arXiv2025

MLLM-Driven Semantic Identifier Generation for Generative Cross-Modal Retrieval

Generative cross-modal retrieval, which treats retrieval as a generation task, has emerged as a promising direction with the rise of Multimodal Large Language Models (MLLMs). In this setting, the model responds to a text query by generating an identifier corresponding to the target image. However, existing methods typically rely on manually crafted string IDs, clustering-based labels, or atomic identifiers requiring vocabulary expansion, all of which face challenges in semantic alignment or scalability.To address these limitations, we propose a vocabulary-efficient identifier generation framework that prompts MLLMs to generate Structured Semantic Identifiers from image-caption pairs. These identifiers are composed of concept-level tokens such as objects and actions, naturally aligning with the model's generation space without modifying the tokenizer. Additionally, we introduce a Rationale-Guided Supervision Strategy, prompting the model to produce a one-sentence explanation alongside each identifier serves as an auxiliary supervision signal that improves semantic grounding and reduces hallucinations during training.

cs.IR↗

arXiv2025

ELEC: Efficient Large Language Model-Empowered Click-Through Rate Prediction

Click-through rate (CTR) prediction plays an important role in online advertising systems. On the one hand, traditional CTR prediction models capture the collaborative signals in tabular data via feature interaction modeling, but they lose semantics in text. On the other hand, Large Language Models (LLMs) excel in understanding the context and meaning behind text, but they face challenges in capturing collaborative signals and they have long inference latency. In this paper, we aim to leverage the benefits of both types of models and pursue collaboration, semantics and efficiency. We present ELEC, which is an Efficient LLM-Empowered CTR prediction framework. We first adapt an LLM for the CTR prediction task. In order to leverage the ability of the LLM but simultaneously keep efficiency, we utilize the pseudo-siamese network which contains a gain network and a vanilla network. We inject the high-level representation vector generated by the LLM into a collaborative CTR model to form the gain network such that it can take advantage of both tabular modeling and textual modeling. However, its reliance on the LLM limits its efficiency. We then distill the knowledge from the gain network to the vanilla network on both the score level and the representation level, such that the vanilla network takes only tabular data as input, but can still generate comparable performance as the gain network. Our approach is model-agnostic. It allows for the integration with various existing LLMs and collaborative CTR models. Experiments on real-world datasets demonstrate the effectiveness and efficiency of ELEC for CTR prediction.

cs.IR↗

arXiv2025

TIDE: Achieving Balanced Subject-Driven Image Generation via Target-Instructed Diffusion Enhancement

Subject-driven image generation (SDIG) aims to manipulate specific subjects within images while adhering to textual instructions, a task crucial for advancing text-to-image diffusion models. SDIG requires reconciling the tension between maintaining subject identity and complying with dynamic edit instructions, a challenge inadequately addressed by existing methods. In this paper, we introduce the Target-Instructed Diffusion Enhancing (TIDE) framework, which resolves this tension through target supervision and preference learning without test-time fine-tuning. TIDE pioneers target-supervised triplet alignment, modelling subject adaptation dynamics using a (reference image, instruction, target images) triplet. This approach leverages the Direct Subject Diffusion (DSD) objective, training the model with paired "winning" (balanced preservation-compliance) and "losing" (distorted) targets, systematically generated and evaluated via quantitative metrics. This enables implicit reward modelling for optimal preservation-compliance balance. Experimental results on standard benchmarks demonstrate TIDE's superior performance in generating subject-faithful outputs while maintaining instruction compliance, outperforming baseline methods across multiple quantitative metrics. TIDE's versatility is further evidenced by its successful application to diverse tasks, including structural-conditioned generation, image-to-image generation, and text-image interpolation. Our code is available at https://github.com/KomJay520/TIDE.

cs.CV↗

arXiv2025

Free Extension of Topological States via Double-zero-index Media

Topological states, known for their robustness against disorder, offer promising avenues for disorder-resistant devices. However, their intrinsic spatial confinement at interfaces imposes geometric constraints that limit the scalability of topological functionalities. Here, we propose a strategy to overcome this limitation by using double-zero-index media to expand topological interfaces. Although occupying finite space, these media are optically equivalent to infinitesimal points, effectively altering the geometry of topological interfaces and breaking conventional bulk-edge correspondence. This strategy enables the spatial expansion of uniform topological states beyond their native interface, offering new possibilities for topological photonic devices. We have verified this behavior through numerical simulations and microwave experiments in a two-dimensional photonic Su-Schrieffer-Heeger lattice. Our findings offer a universal framework to overcome the inherent dimensional limitations of topological states, with implications extending to general wave systems such as acoustic metamaterials.

physics.optics↗

arXiv2024

The persistent Laplacian of non-branching complexes

Non-branching matrices are real matrices with entries in $\{-1,0,1\}$, where each row contains at most two non-zero entries. Such matrices naturally arise in the study of Laplacians of pseudomanifolds and cubical complexes. We show that a basis for the kernel of a non-branching matrix can be computed in near-linear time. Especially, the basis has the special property that the supports of all the column vectors in it are disjoint with each other. Building on this result, we show that the up persistent Laplacian can be computed in near-linear time for a pair of such spaces and its eigenvalues can be more efficiently computed via computing singular values. In addition to that, we analyze the arithmetic operations of up persistent Laplacian with respect to a non-branching filtration. Furthermore, we show that the up persistent Laplacian of $q$-non-branching simplicial complexes can be represented as the Laplacian of an associated hypergraph, thus providing a higher-dimensional generalization of the Kron reduction, as well as a Cheeger-type inequality. Finally, we highlight the efficiency of our method on image data.

math.AT↗

arXiv2024

FedUD: Exploiting Unaligned Data for Cross-Platform Federated Click-Through Rate Prediction

Click-through rate (CTR) prediction plays an important role in online advertising platforms. Most existing methods use data from the advertising platform itself for CTR prediction. As user behaviors also exist on many other platforms, e.g., media platforms, it is beneficial to further exploit such complementary information for better modeling user interest and for improving CTR prediction performance. However, due to privacy concerns, data from different platforms cannot be uploaded to a server for centralized model training. Vertical federated learning (VFL) provides a possible solution which is able to keep the raw data on respective participating parties and learn a collaborative model in a privacy-preserving way. However, traditional VFL methods only utilize aligned data with common keys across parties, which strongly restricts their application scope. In this paper, we propose FedUD, which is able to exploit unaligned data, in addition to aligned data, for more accurate federated CTR prediction. FedUD contains two steps. In the first step, FedUD utilizes aligned data across parties like traditional VFL, but it additionally includes a knowledge distillation module. This module distills useful knowledge from the guest party's high-level representations and guides the learning of a representation transfer network. In the second step, FedUD applies the learned knowledge to enrich the representations of the host party's unaligned data such that both aligned and unaligned data can contribute to federated model training. Experiments on two real-world datasets demonstrate the superior performance of FedUD for federated CTR prediction.

cs.IR↗

arXiv2024

Elucidating the Role of Stacking Faults in TlGaSe$_{2}$ on its Thermoelectric Properties

Thermoelectric materials are of great interest for heat energy harvesting applications. One such promising material is TlGaSe$_{2}$, a p-type semiconducting ternary chalcogenide. Recent reports show it can be processed as a thin film, opening the door for large-scale commercialization. However, TlGaSe$_{2}$ is prone to stacking faults along the [001] stacking direction and their role in its thermoelectric properties has not been understood to date. Herein, TlGaSe$_{2}$ is investigated via (scanning) transmission electron microscopy and first-principles calculations. Stacking faults are found to be present throughout the material, as density functional theory calculations reveal a lack of preferential stacking order. Electron transport calculations show an enhancement of thermoelectric power factors when stacking faults are present. This implies the presence of stacking faults is key to the material's excellent thermoelectric properties along the [001] stacking direction, which can be further enhanced by doping the material to hole carrier concentrations to approx. 10$^{19}$ cm$^{-3}$.

cond-mat.mtrl-sci↗

arXiv2024

Linguistics from a topological viewpoint

Typological databases in linguistics are usually categorical-valued. As a result, it is difficult to have a clear visualization of the data. In this paper, we describe a workflow to analyze the topological shapes of South American languages by applying multiple correspondence analysis technique and topological data analysis methods.

cs.CL↗

arXiv2023

Tunable interfacial chemisorption with atomic-level precision in a graphene WSe2 heterostructure

It has long been an ultimate goal to introduce chemical doping at the atomic level to precisely tune properties of materials. Two-dimensional materials have natural advantage because of its highly-exposed surface atoms, however, it is still a grand challenge to achieve this goal experimentally. Here, we demonstrate the ability to introduce chemical doping in graphene with atomic-level precision by controlling chemical adsorption of individual Se atoms, which are extracted from the underneath WSe2, at the interface of graphene-WSe2 heterostructures. Our scanning tunneling microscopy (STM) measurements, combined with first-principles calculations, reveal that individual Se atoms can chemisorbed on three possible positions in graphene, which generate distinct pseudospin-mediated atomic-scale vortices in graphene. We demonstrate that the chemisorbed positions of individual Se atoms can be manipulated by STM tip, which enables us to achieve atomic-scale controlling quantum interference of the pseudospin-mediated vortices in graphene. This result offers the promise of controlling properties of materials through chemical doping with atomic-level precision.

cond-mat.mtrl-sci↗

arXiv2023

Efficient Bottom-Up Synthesis for Programs with Local Variables

We propose a new synthesis algorithm that can efficiently search programs with local variables (e.g., those introduced by lambdas). Prior bottom-up synthesis algorithms are not able to evaluate programs with free local variables, and therefore cannot effectively reduce the search space of such programs (e.g., using standard observational equivalence reduction techniques), making synthesis slow. Our algorithm can reduce the space of programs with local variables. The key idea, dubbed lifted interpretation, is to lift up the program interpretation process, from evaluating one program at a time to simultaneously evaluating all programs from a grammar. Lifted interpretation provides a mechanism to systematically enumerate all binding contexts for local variables, thereby enabling us to evaluate and reduce the space of programs with local variables. Our ideas are instantiated in the domain of web automation. The resulting tool, Arborist, can automate a significantly broader range of challenging tasks more efficiently than state-of-the-art techniques including WebRobot and Helena.

cs.PL↗