arXiv subjects

Xi Zhu

Publications and source records attributed to Xi Zhu.

At least 55 records · Page 3Linked to original sources

arXiv2024

Knowledge Graph Pruning for Recommendation

Recent years have witnessed the prosperity of knowledge graph based recommendation system (KGRS), which enriches the representation of users, items, and entities by structural knowledge with striking improvement. Nevertheless, its unaffordable computational cost still limits researchers from exploring more sophisticated models. We observe that the bottleneck for training efficiency arises from the knowledge graph, which is plagued by the well-known issue of knowledge explosion. Recently, some works have attempted to slim the inflated KG via summarization techniques. However, these summarized nodes may ignore the collaborative signals and deviate from the facts that nodes in knowledge graph represent symbolic abstractions of entities from the real-world. To this end, in this paper, we propose a novel approach called KGTrimmer for knowledge graph pruning tailored for recommendation, to remove the unessential nodes while minimizing performance degradation. Specifically, we design an importance evaluator from a dual-view perspective. For the collective view, we embrace the idea of collective intelligence by extracting community consensus based on abundant collaborative signals, i.e. nodes are considered important if they attract attention of numerous users. For the holistic view, we learn a global mask to identify the valueless nodes from their inherent properties or overall popularity. Next, we build an end-to-end importance-aware graph neural network, which injects filtered knowledge to enhance the distillation of valuable user-item collaborative signals. Ultimately, we generate a pruned knowledge graph with lightweight, stable, and robust properties to facilitate the following-up recommendation task. Extensive experiments are conducted on three publicly available datasets to prove the effectiveness and generalization ability of KGTrimmer.

cs.IR↗

arXiv2024

When Box Meets Graph Neural Network in Tag-aware Recommendation

Last year has witnessed the re-flourishment of tag-aware recommender systems supported by the LLM-enriched tags. Unfortunately, though large efforts have been made, current solutions may fail to describe the diversity and uncertainty inherent in user preferences with only tag-driven profiles. Recently, with the development of geometry-based techniques, e.g., box embedding, diversity of user preferences now could be fully modeled as the range within a box in high dimension space. However, defect still exists as these approaches are incapable of capturing high-order neighbor signals, i.e., semantic-rich multi-hop relations within the user-tag-item tripartite graph, which severely limits the effectiveness of user modeling. To deal with this challenge, in this paper, we propose a novel algorithm, called BoxGNN, to perform the message aggregation via combination of logical operations, thereby incorporating high-order signals. Specifically, we first embed users, items, and tags as hyper-boxes rather than simple points in the representation space, and define two logical operations to facilitate the subsequent process. Next, we perform the message aggregation mechanism via the combination of logical operations, to obtain the corresponding high-order box representations. Finally, we adopt a volume-based learning objective with Gumbel smoothing techniques to refine the representation of boxes. Extensive experiments on two publicly available datasets and one LLM-enhanced e-commerce dataset have validated the superiority of BoxGNN compared with various state-of-the-art baselines. The code is released online

cs.IR↗

arXiv2024

Efficient Model Compression for Hierarchical Federated Learning

Federated learning (FL), as an emerging collaborative learning paradigm, has garnered significant attention due to its capacity to preserve privacy within distributed learning systems. In these systems, clients collaboratively train a unified neural network model using their local datasets and share model parameters rather than raw data, enhancing privacy. Predominantly, FL systems are designed for mobile and edge computing environments where training typically occurs over wireless networks. Consequently, as model sizes increase, the conventional FL frameworks increasingly consume substantial communication resources. To address this challenge and improve communication efficiency, this paper introduces a novel hierarchical FL framework that integrates the benefits of clustered FL and model compression. We present an adaptive clustering algorithm that identifies a core client and dynamically organizes clients into clusters. Furthermore, to enhance transmission efficiency, each core client implements a local aggregation with compression (LC aggregation) algorithm after collecting compressed models from other clients within the same cluster. Simulation results affirm that our proposed algorithms not only maintain comparable predictive accuracy but also significantly reduce energy consumption relative to existing FL mechanisms.

cs.LG↗

arXiv2024

DynLLM: When Large Language Models Meet Dynamic Graph Recommendation

Last year has witnessed the considerable interest of Large Language Models (LLMs) for their potential applications in recommender systems, which may mitigate the persistent issue of data sparsity. Though large efforts have been made for user-item graph augmentation with better graph-based recommendation performance, they may fail to deal with the dynamic graph recommendation task, which involves both structural and temporal graph dynamics with inherent complexity in processing time-evolving data. To bridge this gap, in this paper, we propose a novel framework, called DynLLM, to deal with the dynamic graph recommendation task with LLMs. Specifically, DynLLM harnesses the power of LLMs to generate multi-faceted user profiles based on the rich textual features of historical purchase records, including crowd segments, personal interests, preferred categories, and favored brands, which in turn supplement and enrich the underlying relationships between users and items. Along this line, to fuse the multi-faceted profiles with temporal graph embedding, we engage LLMs to derive corresponding profile embeddings, and further employ a distilled attention mechanism to refine the LLM-generated profile embeddings for alleviating noisy signals, while also assessing and adjusting the relevance of each distilled facet embedding for seamless integration with temporal graph embedding from continuous time dynamic graphs (CTDGs). Extensive experiments on two real e-commerce datasets have validated the superior improvements of DynLLM over a wide range of state-of-the-art baseline methods.

cs.IR↗

arXiv2024

BezierFormer: A Unified Architecture for 2D and 3D Lane Detection

Lane detection has made significant progress in recent years, but there is not a unified architecture for its two sub-tasks: 2D lane detection and 3D lane detection. To fill this gap, we introduce BézierFormer, a unified 2D and 3D lane detection architecture based on Bézier curve lane representation. BézierFormer formulate queries as Bézier control points and incorporate a novel Bézier curve attention mechanism. This attention mechanism enables comprehensive and accurate feature extraction for slender lane curves via sampling and fusing multiple reference points on each curve. In addition, we propose a novel Chamfer IoU-based loss which is more suitable for the Bézier control points regression. The state-of-the-art performance of BézierFormer on widely-used 2D and 3D lane detection benchmarks verifies its effectiveness and suggests the worthiness of further exploration.

cs.CV↗

arXiv2024

REALM: RAG-Driven Enhancement of Multimodal Electronic Health Records Analysis via Large Language Models

The integration of multimodal Electronic Health Records (EHR) data has significantly improved clinical predictive capabilities. Leveraging clinical notes and multivariate time-series EHR, existing models often lack the medical context relevent to clinical tasks, prompting the incorporation of external knowledge, particularly from the knowledge graph (KG). Previous approaches with KG knowledge have primarily focused on structured knowledge extraction, neglecting unstructured data modalities and semantic high dimensional medical knowledge. In response, we propose REALM, a Retrieval-Augmented Generation (RAG) driven framework to enhance multimodal EHR representations that address these limitations. Firstly, we apply Large Language Model (LLM) to encode long context clinical notes and GRU model to encode time-series EHR data. Secondly, we prompt LLM to extract task-relevant medical entities and match entities in professionally labeled external knowledge graph (PrimeKG) with corresponding medical knowledge. By matching and aligning with clinical standards, our framework eliminates hallucinations and ensures consistency. Lastly, we propose an adaptive multimodal fusion network to integrate extracted knowledge with multimodal EHR data. Our extensive experiments on MIMIC-III mortality and readmission tasks showcase the superior performance of our REALM framework over baselines, emphasizing the effectiveness of each module. REALM framework contributes to refining the use of multimodal EHR data in healthcare and bridging the gap with nuanced medical context essential for informed clinical predictions.

cs.AI↗

arXiv2023

Model Selection for Exposure-Mediator Interaction

In mediation analysis, the exposure often influences the mediating effect, i.e., there is an interaction between exposure and mediator on the dependent variable. When the mediator is high-dimensional, it is necessary to identify non-zero mediators (M) and exposure-by-mediator (X-by-M) interactions. Although several high-dimensional mediation methods can naturally handle X-by-M interactions, research is scarce in preserving the underlying hierarchical structure between the main effects and the interactions. To fill the knowledge gap, we develop the XMInt procedure to select M and X-by-M interactions in the high-dimensional mediators setting while preserving the hierarchical structure. Our proposed method employs a sequential regularization-based forward-selection approach to identify the mediators and their hierarchically preserved interaction with exposure. Our numerical experiments showed promising selection results. Further, we applied our method to ADNI morphological data and examined the role of cortical thickness and subcortical volumes on the effect of amyloid-beta accumulation on cognitive performance, which could be helpful in understanding the brain compensation mechanism.

stat.ME↗

arXiv2023

A survey on deep learning approaches for data integration in autonomous driving system

The perception module of self-driving vehicles relies on a multi-sensor system to understand its environment. Recent advancements in deep learning have led to the rapid development of approaches that integrate multi-sensory measurements to enhance perception capabilities. This paper surveys the latest deep learning integration techniques applied to the perception module in autonomous driving systems, categorizing integration approaches based on "what, how, and when to integrate". A new taxonomy of integration is proposed, based on three dimensions: multi-view, multi-modality, and multi-frame. The integration operations and their pros and cons are summarized, providing new insights into the properties of an "ideal" data integration approach that can alleviate the limitations of existing methods. After reviewing hundreds of relevant papers, this survey concludes with a discussion of the key features of an optimal data integration approach.

cs.RO↗

arXiv2023

NeMO: Neural Map Growing System for Spatiotemporal Fusion in Bird's-Eye-View and BDD-Map Benchmark

Vision-centric Bird's-Eye View (BEV) representation is essential for autonomous driving systems (ADS). Multi-frame temporal fusion which leverages historical information has been demonstrated to provide more comprehensive perception results. While most research focuses on ego-centric maps of fixed settings, long-range local map generation remains less explored. This work outlines a new paradigm, named NeMO, for generating local maps through the utilization of a readable and writable big map, a learning-based fusion module, and an interaction mechanism between the two. With an assumption that the feature distribution of all BEV grids follows an identical pattern, we adopt a shared-weight neural network for all grids to update the big map. This paradigm supports the fusion of longer time series and the generation of long-range BEV local maps. Furthermore, we release BDD-Map, a BDD100K-based dataset incorporating map element annotations, including lane lines, boundaries, and pedestrian crossing. Experiments on the NuScenes and BDD-Map datasets demonstrate that NeMO outperforms state-of-the-art map segmentation methods. We also provide a new scene-level BEV map evaluation setting along with the corresponding baseline for a more comprehensive comparison.

cs.CV↗

arXiv2023

Anchor Free remote sensing detector based on solving discrete polar coordinate equation

As the rapid development of depth learning, object detection in aviatic remote sensing images has become increasingly popular in recent years. Most of the current Anchor Free detectors based on key point detection sampling directly regression and classification features, with the design of object loss function based on the horizontal bounding box. It is more challenging for complex and diverse aviatic remote sensing object. In this paper, we propose an Anchor Free aviatic remote sensing object detector (BWP-Det) to detect rotating and multi-scale object. Specifically, we design a interactive double-branch(IDB) up-sampling network, in which one branch gradually up-sampling is used for the prediction of Heatmap, and the other branch is used for the regression of boundary box parameters. We improve a weighted multi-scale convolution (WmConv) in order to highlight the difference between foreground and background. We extracted Pixel level attention features from the middle layer to guide the two branches to pay attention to effective object information in the sampling process. Finally, referring to the calculation idea of horizontal IoU, we design a rotating IoU based on the split polar coordinate plane, namely JIoU, which is expressed as the intersection ratio following discretization of the inner ellipse of the rotating bounding box, to solve the correlation between angle and side length in the regression process of the rotating bounding box. Ultimately, BWP-Det, our experiments on DOTA, UCAS-AOD and NWPU VHR-10 datasets show, achieves advanced performance with simpler models and fewer regression parameters.

cs.CV↗

arXiv2021

Unraveling the Excitonic Transition and Associated Dynamics in Confined Long Linear Carbon-chains with Time-Resolved Resonance Raman Scattering

Long linear carbon-chains have been attracting intense interest arising from the remarkable properties predicted and their potential applications in future nanotechnology. Here we comprehensively interrogate the excitonic transitions and the associated relaxation dynamics of nanotube confined long linear carbon-chains by using steady state and time-resolved Raman spectroscopies. The exciton relaxation dynamics on the confined carbon-chains occurs on a hundreds of picoseconds timescale, in strong contrast to the host dynamics that occurs on a few picosecond timescale. A prominent time-resolved Raman response is observed over a broad energy range extending from 1.2 to 2.8 eV, which includes the strong Raman resonance region around 2.2 eV. Evidence for a strong coupling between the chain and the nanotube host is found from the dynamics at high excitation energies which provides a clear evidence for an efficient energy transfer from the host carbon nanotube to the chain. Our experimental study presents the first unique characterization of the long linear carbon-chain exciton dynamics, providing indispensable knowledge for the understanding of the interactions between different carbon allotropes.

cond-mat.mtrl-sci↗

arXiv2021

Tensor Network for Supervised Learning at Finite Temperature

The large variation of datasets is a huge barrier for image classification tasks. In this paper, we embraced this observation and introduce the finite temperature tensor network (FTTN), which imports the thermal perturbation into the matrix product states framework by placing all images in an environment with constant temperature, in analog to energy-based learning. Tensor network is chosen since it is the best platform to introduce thermal fluctuation. Different from traditional network structure which directly takes the summation of individual losses as its loss function, FTTN regards it as thermal average loss computed from the entanglement with the environment. The temperature-like parameter can be automatically optimized, which gives each database an individual temperature. FTTN obtains improvement in both test accuracy and convergence speed in several datasets. The non-zero temperature automatically separates similar features, avoiding the wrong classification in previous architecture. The thermal fluctuation may give a better improvement in other frameworks, and we may also implement the temperature of database to improve the training effect.

cs.LG↗

arXiv2021

Effect of Varying the TD-lc-DFTB Range-Separation Parameter on Charge and Energy Transfer in a Model Pentacene/Buckminsterfullerene Heterojunction

Density-functional tight binding (DFTB) has become a popular form of approximate density-functional theory (DFT) based upon a minimal valence basis set and neglect of all but two center integrals. We report the results of our tests of a recent long-range correction (lc) for time-dependent (TD) lc-DFTB by carrying out TD-lc-DFTB fewest switches surface hopping (FSSH) calculations of energy and charge transfer times using the relatively new DFTBaby program. An advantage of this method is the ability to run enough trajectories to get meaningful ensemble averages. Our interest in the present work is less in determining exact energy and charge transfer rates than in understanding how the results of these calculations vary with the value of the range-separation parameter (Rlc = 1/μ) for a model organic solar cell heterojunction consisting of a van der Waals complex P/F made up of single pentacene (P) molecule together with a single buckminsterfullerene (F) molecule. The default value of Rlc = 3.03 a0 is found to be much too small as neither energy nor charge transfer is observed until Rlc ~ 10 a0. Tests at a single geometry show that best agreement with high-quality ab-initio spectra is obtained in the limit of no lc (i.e., very large Rlc.) A plot of energy and charge transfer rates as a function of Rlc is provided which suggests that a value of Rlc ~ 15 a0 yields the typical literature charge transfer time of about 100 fs. However, energy and charge transfer times become as high as ~ 300 fs for Rlc ~ 25 a0. A closer examination of the charge transfer process P*/F to P+/F- shows that the initial electron transfer is accompanied by a partial delocalization of the P hole onto F which then relocalizes back onto P, consistent with a polaron-like picture in which the nuclei relax to stabilize the resultant redistribution of charges.

physics.chem-ph↗

arXiv2020

Overcoming Language Priors with Self-supervised Learning for Visual Question Answering

Most Visual Question Answering (VQA) models suffer from the language prior problem, which is caused by inherent data biases. Specifically, VQA models tend to answer questions (e.g., what color is the banana?) based on the high-frequency answers (e.g., yellow) ignoring image contents. Existing approaches tackle this problem by creating delicate models or introducing additional visual annotations to reduce question dependency while strengthening image dependency. However, they are still subject to the language prior problem since the data biases have not been even alleviated. In this paper, we introduce a self-supervised learning framework to solve this problem. Concretely, we first automatically generate labeled data to balance the biased data, and propose a self-supervised auxiliary task to utilize the balanced data to assist the base VQA model to overcome language priors. Our method can compensate for the data biases by generating balanced data without introducing external annotations. Experimental results show that our method can significantly outperform the state-of-the-art, improving the overall accuracy from 49.50% to 57.59% on the most commonly used benchmark VQA-CP v2. In other words, we can increase the performance of annotation-based methods by 16% without using external annotations.

cs.CV↗

arXiv2019

Symmetrical Graph Neural Network for Quantum Chemistry, with Dual R/K Space

Most of current neural network models in quantum chemistry (QC) exclude the molecular symmetry, separate the well-correlated real space (R space), and momenta space (K space) into two individuals, which lack the essential physics in molecular chemistry. In this work, by endorsing the molecular symmetry and elementals of group theory, we propose a comprehensive method to apply symmetry in the graph neural network (SY-GNN), which extends the property-predicting coverage to all the orbital symmetry for both ground and excited states. SY-GNN shows excellent performance in predicting both the absolute and relative of R and K spaces quantities. Besides the numerical properties, SY-GNN also can predict the orbitals distributions in real space, providing the active regions of chemical reactions. We believe the symmetry endorsed deep learning scheme covers the significant physics inside and is essential for the application of neural networks in QC and many other research fields in the future.

physics.chem-ph↗

arXiv2019

Fixed-Size Ordinally Forgetting Encoding Based Word Sense Disambiguation

In this paper, we present our method of using fixed-size ordinally forgetting encoding (FOFE) to solve the word sense disambiguation (WSD) problem. FOFE enables us to encode variable-length sequence of words into a theoretically unique fixed-size representation that can be fed into a feed forward neural network (FFNN), while keeping the positional information between words. In our method, a FOFE-based FFNN is used to train a pseudo language model over unlabelled corpus, then the pre-trained language model is capable of abstracting the surrounding context of polyseme instances in labelled corpus into context embeddings. Next, we take advantage of these context embeddings towards WSD classification. We conducted experiments on several WSD data sets, which demonstrates that our proposed method can achieve comparable performance to that of the state-of-the-art approach at the expense of much lower computational cost.

cs.CL↗

arXiv2018

The Newton Scheme for Deep Learning

We introduce a neural network (NN) strictly governed by Newton's Law, with the nature required basis functions derived from the fundamental classic mechanics. Then, by classifying the training model as a quick procedure of 'force pattern' recognition, we developed the Newton physics-based NS scheme. Once the force pattern is confirmed, the neuro network simply does the checking of the 'pattern stability' instead of the continuous fitting by computational resource consuming big data-driven processing. In the given physics's law system, once the field is confirmed, the mathematics bases for the force field description actually are not diverged but denumerable, which can save the function representations from the exhaustible available mathematics bases. In this work, we endorsed Newton's Law into the deep learning technology and proposed Newton Scheme (NS). Under NS, the user first identifies the path pattern, like the constant acceleration movement.The object recognition technology first loads mass information, then, the NS finds the matched physical pattern and describe and predict the trajectory of the movements with nearly zero error. We compare the major contribution of this NS with the TCN, GRU and other physics inspired 'FIND-PDE' methods to demonstrate fundamental and extended applications of how the NS works for the free-falling, pendulum and curve soccer balls.The NS methodology provides more opportunity for the future deep learning advances.

cs.LG↗

arXiv2017

The Evolution of Alq3 Films Exposure to Air

Small molecular solar cell becomes more stable when a thin tris-(8-hydroxyquinoline) aluminum (Alq3) buffer layer instead of bathocuproine (BCP) is inserted between the active layer and electrode. In this work, we introduce a single layer device (ITO / Alq3 / Au) exposure to air to investigate the role of Alq3 in organic solar cells. The large PL intensity and undetectable Raman peaks of 8-hydroxyquinoline (8-Hq), a degradation product of Alq3 through chemical reaction, indicate that the degradation of Alq3 is not severe even after a few hundred hours exposure. Thus the Alq3 buffer layer gives a good protection of the active layers against oxygen\water permeation comparing with BCP. However, the small quantity of degraded Alq3 products can enhance the current density (after 15 h) in single layer Alq3 device, which is reflected in the decrease of parallel resistance Rp. Therefore, due to the enhanced charge transport and the long term better protection from ambient oxygen/water, the lifetime of small molecular solar cell with Alq3 buffer layer is improved.

cond-mat.mtrl-sci↗