arXiv subjects

Yuhong Chen

Publications and source records attributed to Yuhong Chen.

13 recordsLinked to original sources

arXiv2026

Rethinking Message Passing as Retrieval for Text-Attributed Graph Learning

Graph neural networks (GNNs) are typically conceptualized as message-passing neural networks, yet it remains unclear why neighborhood aggregation reliably outperforms node-wise multilayer perceptrons (MLPs). Despite its empirical success, this paradigm can be computationally expensive and sensitive to imperfect graph structures. In this work, we present a retrieval-augmented view of GNNs: each layer makes predictions by applying an MLP to a node representation together with a permutation-invariant summary of retrieved graph context. Motivated by this perspective, we propose RTA, a simple MLP-based framework that replaces structural message passing with label-aware retrieval and propagation. We provide theoretical insights that (i) connect retrieval-based aggregation to softmax-attention message passing, and (ii) establish the robustness of retrieved-context supervision to mis-retrieved outliers. Experiments on multiple text-attributed graph benchmarks show that RTA matches or even outperforms strong GNN and graph LLM baselines while improving efficiency and robustness across diverse scenarios.

cs.LG↗

arXiv2026

CodeHID: Learning an Addressable Hierarchical Code Index for Generative Code Retrieval

Code retrieval models have predominantly relied on a flat matching paradigm that treats code snippets as independent candidates, making them less capable of distinguishing similar code candidates. Generative retrieval offers a solution by constructing a learnable index over the code corpus, guiding the retriever to better understand how code candidates are semantically organized and addressed. However, naively applying generative retrieval in the code retrieval task may result in operating over an identifier space whose prefixes do not correspond to meaningful code-semantic regions. In this paper, we propose CodeHID, a generative code retrieval framework that reformulates the code retrieval task from flat candidate matching to coarse-to-fine semantic address generation. CodeHID relies on two core components. First, Pseudo-Neighbor Guided DocID Learning constructs a globally static hierarchical index by applying multi-level residual quantization and $k$-nearest-neighbor pseudo-supervision, ensuring that semantically related code snippets share prefixes while preserving target-level separability. Second, Dual-Phase DocID Generation Guidance reliably navigates this fixed index by combining training-side ranking enhancements, using hard negatives and rank distillation, with inference-side candidate constraints and prefix-aware decoding. Extensive experiments on CoSQA and ProCQA benchmarks demonstrate that CodeHID outperforms existing sparse retrieval, pre-trained code models, dense code retrieval, and generative retrieval baselines by a large margin in most cases, achieving particularly strong improvements in rank-one retrieval metrics.

cs.IR↗

arXiv2026

Knowledge-Geometry Decoupling: Refreshable Pretrained Transfer for Streaming Recommendation

Industrial recommenders increasingly adopt the pretrain-then-transfer paradigm, yet behavioral distribution drift raises two questions: what to learn from behavior sequences, and how to transfer the learned knowledge while the pretrained model is continually refreshed. To resolve them, we propose Knowledge-Geometry Decoupling (KGD). For what to learn, conventional next-token prediction treats adjacency as dependency and may encode spurious transitions across unrelated sessions. We introduce Behavioral Multi-Token Prediction (BMTP) to retain only collaboratively or semantically related future items as supervision, yielding cleaner and more transferable behavioral knowledge. For how to transfer, pretrained knowledge and task-specific geometry impose conflicting optimization demands on shared parameters. To handle it, KGD assigns them to separate parameter sets: a refreshable encoder owns behavioral knowledge, while a task learner reads contextualized encoder states through read-only cross-attention and writes task-specific geometry through Anchored Calibration Residual (ACR) orthogonal to the pretrained embedding. The decoupled ownership enables continual knowledge refresh without task-gradient interference or invalidating downstream adaptation. KGD improves over strong pretrain-transfer baselines by 4-12% on eight public benchmarks and sustains its advantage over a 90-day production stream where baselines show no gains. KGD has been fully deployed in Shopee. In a live A/B test on Shopee Homepage Search, it increases GMV per user by 1.75% and advertising revenue by 1.53%, demonstrating its high practical value. We provide the core implementation of KGD at https://github.com/FuCongResearchSquad/KGD4REC.

cs.IR↗

arXiv2026

Edge-Aware Curvature Modeling for Graph Understanding in Large Language Models

Recently, graph-aware Large Language Models (LLMs) have shown promising capabilities in jointly modeling graph-structured data and textual information. Existing approaches typically employ a graph encoder and a frozen LLM to obtain node representations from graph and textual views, followed by node-level alignment to bridge the two modalities. However, such alignment mechanisms primarily focus on node information while overlooking edge-level structures, leading to suboptimal information propagation across views. In this work, we conduct a comprehensive theoretical analysis to uncover why node-level alignment is insufficient for aligning textual and graph representations. Specifically, we prove theoretically for the first time that neglecting edge information leads to suboptimal solutions and negatively curved edges induce bottlenecked information flow, giving rise to the over-squashing phenomenon between graph and textual views. To address the two challenges, we innovatively proposed a CureLLM framework of Curvature-enhanced Graph Representations for Large Language Model whose goal is to inject the signals of edge information into the existing LLMs. Specifically, CureLLM first introduces the training-free textual prompt mechanism to make the LLM model generate the output directly based on the edge-aware prompt without learnable parameter costs. Furthermore, a novel curvature-aware graph representation learning is designed to capture the edge structure information to enhance the downstream tasks, where the message passing between text and graph representations only depends on edges with positive curvature. Finally, we conduct evaluations with 20 different compared methods on 11 real world datasets from various domains and the experiment results demonstrate the superiority of our proposed CureLLM framework.

cs.IR↗

arXiv2026

Enzyme: Incremental View Maintenance for Data Engineering

Materialized views are a core construct in database systems, used to accelerate analytical queries and optimize batch pipelines for extract-transform-load (ETL) workflows. Maintaining view consistency as underlying data evolves is a fundamental challenge, especially in high-throughput and real-time settings. Incremental view maintenance (IVM) has been studied for decades and continues to attract significant investment from major database vendors. However, most industrial systems either offer limited SQL-operator coverage or require users to hand-tune refresh strategies. This paper presents Enzyme, an IVM engine developed at Databricks to power Spark Declarative Pipelines. It provides a built-in, end-to-end approach to incremental pipelines, utilizing materialized views as first-class building blocks. By automating refresh planning, Enzyme reduces total cost of ownership and lets users focus on business logic rather than MV mechanics. Validation across thousands of large-scale production pipelines spanning diverse application domains has demonstrated substantial computational efficiency gains, yielding a cumulative daily compute reduction of billions of CPU seconds. Built atop Apache Spark primitives, Enzyme adds a cost-based optimization layer that selects refresh strategies for collections of materialized views organized into pipelines. Enzyme's modular architecture is designed to generalize across data sources and query engines. We present key design decisions for incremental refresh planning and execution, including optimizations that exploit batching opportunities across materialized view sources. Experimental results on standard benchmarks demonstrate significant performance improvements at scale.

cs.DB↗

arXiv2026

SWE-AGILE: A Software Agent Framework for Efficiently Managing Dynamic Reasoning Context

Prior representative ReAct-style approaches in autonomous Software Engineering (SWE) typically lack the explicit System-2 reasoning required for deep analysis and handling complex edge cases. While recent reasoning models demonstrate the potential of extended Chain-of-Thought (CoT), applying them to the multi-turn SWE task creates a fundamental dilemma: retaining full reasoning history leads to context explosion and ``Lost-in-the-Middle'' degradation, while discarding it would force the agent to redundantly re-reason at every step. To address these challenges, we propose SWE-AGILE, a novel software agent framework designed to bridge the gap between reasoning depth, efficiency, and context constraints. SWE-AGILE introduces a Dynamic Reasoning Context strategy, maintaining a ``sliding window'' of detailed reasoning for immediate continuity to prevent redundant re-analyzing, while compressing historical reasoning content into concise Reasoning Digests. Empirically, SWE-AGILE sets a new standard for 7B-8B models on SWE-Bench-Verified using only 2.2k trajectories and 896 tasks. Code is available at https://github.com/KDEGroup/SWE-AGILE.

cs.AI↗

arXiv2025

2D MoS$_2$/Au interfaces for enhanced opto-electronic response with sub-bandgap photons

Monolayer MoS$_2$ is a direct band gap semiconductor with potential applications in optoelectronics and photonics. MoS$_2$ also has a large optical nonlinearity. However, the atomic thickness of the monolayer limits the strength of the measured functional signals, such as the photocurrent or photoluminescence, in optoelectronic devices. Here, we show that photocurrent in monolayer MoS$_2$ can be induced by sub-band gap photons by depositing Au nanoparticles on it. In this system, the nonlinear light-matter interaction in Au nanoparticles enhanced by the localized surface plasmons results in the generation of supercontinuum, which is reabsorbed by MoS$_2$ due to efficient resonant energy transfer. Au nanoparticle assisted photocurrent is more than an order of magnitude larger than two-photon photocurrent in monolayer MoS$_2$. Optimization of the shape, size and composition of the nanoparticle has the potential to enhance the photocurrent significantly with the prospect of applications in the detection of NIR photons, and related technologies including optical telecommunication.

physics.optics↗

arXiv2025

Coevolution of Opinion Dynamics and Recommendation System: Modeling, Analysis and Reinforcement Learning Based Manipulation

In this work, we develop an analytical framework that integrates opinion dynamics with a recommendation system. By incorporating elements such as collaborative filtering, we provide a precise characterization of how recommendation systems shape interpersonal interactions and influence opinion formation. Moreover, the property of the coevolution of both opinion dynamics and recommendation systems is also shown. Specifically, the convergence of this coevolutionary system is theoretically proved, and the mechanisms behind filter bubble formation are elucidated. Our analysis of the maximum number of opinion clusters shows how recommendation system parameters affect opinion grouping and polarization. Additionally, we incorporate the influence of propagators into our model and propose a reinforcement learning-based solution. The analysis and the propagation solution are demonstrated in simulations using the Yelp data set.

eess.SY↗

arXiv2024

Multi-View Incremental Learning with Structured Hebbian Plasticity for Enhanced Fusion Efficiency

The rapid evolution of multimedia technology has revolutionized human perception, paving the way for multi-view learning. However, traditional multi-view learning approaches are tailored for scenarios with fixed data views, falling short of emulating the intricate cognitive procedures of the human brain processing signals sequentially. Our cerebral architecture seamlessly integrates sequential data through intricate feed-forward and feedback mechanisms. In stark contrast, traditional methods struggle to generalize effectively when confronted with data spanning diverse domains, highlighting the need for innovative strategies that can mimic the brain's adaptability and dynamic integration capabilities. In this paper, we propose a bio-neurologically inspired multi-view incremental framework named MVIL aimed at emulating the brain's fine-grained fusion of sequentially arriving views. MVIL lies two fundamental modules: structured Hebbian plasticity and synaptic partition learning. The structured Hebbian plasticity reshapes the structure of weights to express the high correlation between view representations, facilitating a fine-grained fusion of view representations. Moreover, synaptic partition learning is efficient in alleviating drastic changes in weights and also retaining old knowledge by inhibiting partial synapses. These modules bionically play a central role in reinforcing crucial associations between newly acquired information and existing knowledge repositories, thereby enhancing the network's capacity for generalization. Experimental results on six benchmark datasets show MVIL's effectiveness over state-of-the-art methods.

cs.CV↗

arXiv2023

SCP: Scene Completion Pre-training for 3D Object Detection

3D object detection using LiDAR point clouds is a fundamental task in the fields of computer vision, robotics, and autonomous driving. However, existing 3D detectors heavily rely on annotated datasets, which are both time-consuming and prone to errors during the process of labeling 3D bounding boxes. In this paper, we propose a Scene Completion Pre-training (SCP) method to enhance the performance of 3D object detectors with less labeled data. SCP offers three key advantages: (1) Improved initialization of the point cloud model. By completing the scene point clouds, SCP effectively captures the spatial and semantic relationships among objects within urban environments. (2) Elimination of the need for additional datasets. SCP serves as a valuable auxiliary network that does not impose any additional efforts or data requirements on the 3D detectors. (3) Reduction of the amount of labeled data for detection. With the help of SCP, the existing state-of-the-art 3D detectors can achieve comparable performance while only relying on 20% labeled data.

cs.CV↗

arXiv2023

Beyond Graph Convolutional Network: An Interpretable Regularizer-centered Optimization Framework

Graph convolutional networks (GCNs) have been attracting widespread attentions due to their encouraging performance and powerful generalizations. However, few work provide a general view to interpret various GCNs and guide GCNs' designs. In this paper, by revisiting the original GCN, we induce an interpretable regularizer-centerd optimization framework, in which by building appropriate regularizers we can interpret most GCNs, such as APPNP, JKNet, DAGNN, and GNN-LF/HF. Further, under the proposed framework, we devise a dual-regularizer graph convolutional network (dubbed tsGCN) to capture topological and semantic structures from graph data. Since the derived learning rule for tsGCN contains an inverse of a large matrix and thus is time-consuming, we leverage the Woodbury matrix identity and low-rank approximation tricks to successfully decrease the high computational complexity of computing infinite-order graph convolutions. Extensive experiments on eight public datasets demonstrate that tsGCN achieves superior performance against quite a few state-of-the-art competitors w.r.t. classification tasks.

cs.LG↗

arXiv2021

The Breakthrough Listen Search For Intelligent Life Near the Galactic Center I

A line-of-sight towards the Galactic Center (GC) offers the largest number of potentially habitable systems of any direction in the sky. The Breakthrough Listen program is undertaking the most sensitive and deepest targeted SETI surveys towards the GC. Here, we outline our observing strategies with Robert C. Byrd Green Bank Telescope (GBT) and Parkes telescope to conduct 600 hours of deep observations across 0.7--93 GHz. We report preliminary results from our survey for ETI beacons across 1--8 GHz with 7.0 and 11.2 hours of observations with Parkes and GBT, respectively. With our narrowband drifting signal search, we were able to place meaningful constraints on ETI transmitters across 1--4 GHz and 3.9--8 GHz with EIRP limits of $\geq$4$\times$10$^{18}$ W among 60 million stars and $\geq$5$\times$10$^{17}$ W among half a million stars, respectively. For the first time, we were able to constrain the existence of artificially dispersed transient signals across 3.9--8 GHz with EIRP $\geq$1$\times$10$^{14}$ W/Hz with a repetition period $\leq$4.3 hours. We also searched our 11.2 hours of deep observations of the GC and its surrounding region for Fast Radio Burst-like magnetars with the DM up to 5000 pc cm$^{-3}$ with maximum pulse widths up to 90 ms at 6 GHz. We detected several hundred transient bursts from SGR J1745$-$2900, but did not detect any new transient burst with the peak luminosity limit across our observed band of $\geq$10$^{31}$ erg s$^{-1}$ and burst-rate of $\geq$0.23 burst-hr$^{-1}$. These limits are comparable to bright transient emission seen from other Galactic radio-loud magnetars, constraining their presence at the GC.

astro-ph.HE↗

arXiv2020

Distributed Link Removal Strategy for Networked Meta-Population Epidemics and its Application to the Control of the COVID-19 Pandemic

In this paper, we investigate the distributed link removal strategy for networked meta-population epidemics. In particular, a deterministic networked susceptible-infected-recovered (SIR) model is considered to describe the epidemic evolving process. In order to curb the spread of epidemics, we present the spectrum-based optimization problem involving the Perron-Frobenius eigenvalue of the matrix constructed by the network topology and transition rates. A modified distributed link removal strategy is developed such that it can be applied to the SIR model with heterogeneous transition rates on weighted digraphs. The proposed approach is implemented to control the COVID-19 pandemic by using the reported infected and recovered data in each state of Germany. The numerical experiment shows that the infected percentage can be significantly reduced by using the distributed link removal strategy.

physics.soc-ph↗