Article
Computer Science, Information Systems
Yuxun Lu, Kosuke Nakamura, Ryutaro Ichise
Summary: Meta-learning has been proven to be effective for the cold-start problem of recommender systems. However, traditional gradient-based systems raise privacy concerns and require large amounts of user-item interactions. Our proposed HyperRS, a Hypernetwork-based Recommender System, overcomes these limitations and outperforms other state-of-the-art meta-learning recommender systems for the user cold-start problem.
Article
Computer Science, Artificial Intelligence
Yan Zhang, Ivor W. Tsang, Hongzhi Yin, Guowu Yang, Defu Lian, Jingjing Li
Summary: This paper proposes a Deep Pairwise Hashing (DPH) method to map users and items to binary vectors in Hamming space, which efficiently calculates a user's preference for an item. To address data sparsity and cold-start problems, user-item interactive information and item content information are combined to learn effective representations. DPH achieves significant improvement in data sparsity and item cold-start recommendation compared to state-of-the-art frameworks.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2022)
Article
Computer Science, Artificial Intelligence
Renchu Guan, Haoyu Pang, Fausto Giunchiglia, Yanchun Liang, Xiaoyue Feng
Summary: The cold-start problem limits the effectiveness of recommendation systems. There are two main strategies to address this problem: cross-domain recommendation (CDR) and meta-learning. However, CDR methods lack optimization for the few-shot problem, while most meta-learning approaches ignore cross-domain information. Therefore, a novel approach called MetaCDR is proposed, which combines domain knowledge and meta-optimization to solve the cold-start problem.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Computer Science, Information Systems
Faryad Tahmasebi, Majid Meghdadi, Sajad Ahmadian, Khashayar Valiallahi
Summary: This paper introduces a hybrid recommendation method based on profile expansion technique to tackle the cold start problem in recommender systems. By incorporating demographic data and rating data, the proposed method enriches the rating profile of users and improves the system's performance in predicting unseen items.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Junmei Feng, Zhaoqiang Xia, Xiaoyi Feng, Jinye Peng
Summary: Recommender systems aim to predict user demands by analyzing preferences, providing personalized recommendations. A novel CF ranking model is proposed in this paper to tackle the new user cold start problem, combining PMF rating-oriented approach and BPR pairwise ranking-oriented approach effectively.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Tieyun Qian, Yile Liang, Qing Li, Hui Xiong
Summary: Rating prediction is a classic problem addressed by matrix factorization, but recent advancements in deep learning, particularly graph neural networks, have shown impressive progress. This study introduces a new AGNN framework that utilizes attribute graphs to learn preference embeddings for strict cold start users/items.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2022)
Article
Chemistry, Multidisciplinary
Kai Zhuang, Sen Wu, Shuaiqi Liu
Summary: Loan recommendation is beneficial but challenged by cold start and default risk. We propose CSRLoan, a method using pretraining and dual neural matrix factorization to address these challenges and consider semantic information and default risk simultaneously.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Artificial Intelligence
Jingjing Li, Ke Lu, Zi Huang, Heng Tao Shen
Summary: The paper discusses the importance of hits in web systems and the critical role of recommender systems in discovering and displaying interesting items to users. The authors propose a novel approach that addresses both cold-start and long-tail recommendation, tackling the challenges of new users and surprising users.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2021)
Article
Education & Educational Research
Joy Jeevamol, V. G. Renumol
Summary: This paper addresses the new user cold-start problem in e-learning content RSs by proposing an ontology-based content recommender system. By incorporating additional learner data in the recommendation process, the proposed model provides more reliable and personalized recommendations.
EDUCATION AND INFORMATION TECHNOLOGIES
(2021)
Article
Chemistry, Multidisciplinary
Leonor Fernandes, Vera Migueis, Ivo Pereira, Eduardo Oliveira
Summary: This paper introduces a hybrid recommender system that combines four independent systems to improve the accuracy and personalization of recommendations using transactional and portfolio information.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Hardware & Architecture
Syed Irteza Hussain Jafri, Rozaida Ghazali, Irfan Javid, Yana Mazwin Mohmad Hassim, Mubashir Hayat Khan
Summary: Recommender systems are crucial in the digital world and modern economy, providing tailored advice and reducing information overload. To address the cold-start issue, we propose a hybrid deep-learning-based strategy that enriches user and item profiles using collaborative filtering and pretrained deep learning models. By creating more precise and tailored similarity matrices, the proposed method outperforms baseline techniques in terms of precision and rate coverage.
Article
Computer Science, Artificial Intelligence
Zhi Li, Daichi Amagata, Yihong Zhang, Takuya Maekawa, Takahiro Hara, Kei Yonekawa, Mori Kurokawa
Summary: This study focuses on flash sale recommendations and proposes a meta-learning-based recommender system that can handle users' period-specific preferences and the cold-start problem. Experimental results show significant improvements in flash sale recommendations and most of the non-flash sale cold-start recommendations.
KNOWLEDGE-BASED SYSTEMS
(2022)
Review
Chemistry, Multidisciplinary
Nor Aniza Abdullah, Rasheed Abubakar Rasheed, Mohd Hairul Nizam Md. Nasir, Md Mujibur Rahman
Summary: This study explores various methods and challenges researchers face in obtaining auxiliary information for cold start recommendations in recommender systems. Machine learning algorithms are typically used to build prediction models for cold start recommendation, while understanding similar user profiles can serve as auxiliary information.
APPLIED SCIENCES-BASEL
(2021)
Article
Computer Science, Information Systems
Haitao He, Ruixi Zhang, Yangsen Zhang, Jiadong Ren
Summary: In order to solve the user cold-start challenge in multimedia recommender systems, a new model named USBE is proposed in this paper. The model does not require the new user's personal and social information to address the cold-start challenge, and new users can overcome it through a simple system experience. By considering user-similarity and the discrimination of multimedia items, the model recommends suitable items for cold-start users, allowing users to choose and provide feedback independently. The proposed model is lightweight, low delay, and introduces a new cold-start mode. To complement the USBE model, a cyclic training multilayer perceptron model (Re-NN) is proposed to understand the changes in user-similarity for new users. Experiments conducted on the Movielens dataset demonstrate that our model achieves good results and outperforms existing methods after 4 rounds of cold-start recommendations.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Artificial Intelligence
Israr ur Rehman, Waqar Ali, Zahoor Jan, Zulfiqar Ali, Hui Xu, Jie Shao
Summary: The performance of recommendation engines is challenged by cold-start issues. Traditional recommendation techniques have limited capability to address this problem due to data-thirsty machine learning models. On the contrary, the recent success of meta-learning, with its few-shot learning capabilities, has attracted attention for accommodating new tasks with few data samples. We propose a Contextually Augmented Meta-Learning recommender system (CAML) to improve task adaptation capability by augmenting contextual features into a meta-learning model.
Article
Computer Science, Artificial Intelligence
Caoyuan Li, Hong-Bo Xie, Xuhui Fan, Richard Yi Da Xu, Sabine Van Huffel, Kerrie Mengersen
Summary: The article introduces a hierarchical kernelized sparse Bayesian matrix factorization (KSBMF) model, which automatically infers parameters and latent variables, achieving low-rankness and columnwise sparsity through enforced constraints. Experimental results show that KSBMF outperforms state-of-the-art approaches for image restoration tasks.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2021)
Article
Engineering, Electrical & Electronic
Ziyue Zhang, Shuai Jiang, Congzhentao Huang, Richard Yi Da Xu
Summary: This study addresses the challenges of clothing changes and lack of data labels in person ReID, proposing a novel unsupervised model that simultaneously solves both problems. The model includes synthetic augmentation and feature restriction, and outperforms existing methods on clothing change ReID datasets.
IEEE SIGNAL PROCESSING LETTERS
(2022)
Article
Computer Science, Information Systems
Zhenguo Shi, J. Andrew Zhang, Richard Yida Xu, Qingqing Cheng
Summary: Deep Learning plays a crucial role in device-free WiFi Sensing for human activity recognition. However, challenges such as the need for a large amount of training samples and network adaptation to new environments still exist. To address these challenges, we propose a novel scheme using matching network with enhanced channel state information (MatNet-eCSI) for one-shot learning HAR. Our proposed scheme improves and condenses activity-related information in input signals, significantly reducing computational complexity. It also utilizes data from previously seen environments (PSE) for effective training. Experimental results show that our scheme outperforms state-of-the-art HAR methods, achieving higher recognition accuracy and less training time.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2022)
Article
Computer Science, Artificial Intelligence
Chenghao Zhang, Gaofeng Meng, Richard Yi Da Xu, Shiming Xiang, Chunhong Pan
Summary: State-of-the-art stereo matching models trained on synthetic datasets often struggle to generalize to real-world datasets. This study proposes an end-to-end domain alignment stereo network (DAStereo) that focuses on aligning data domains in feature space. The introduction of a domain alignment module (DAM) and adversarial learning significantly improves the model's performance.
Article
Biochemical Research Methods
Xian L. Yang, Shuo J. Wang, Yuting L. Xing, Ling J. Li, Richard Yi Da L. Xu, Karl J. Friston, Yike L. Guo
Summary: In this study, a Bayesian data assimilation framework is proposed for estimating the parameters of COVID-19 transmission dynamics, specifically the instantaneous reproduction number R-t. The system, called DARt, addresses issues such as lagging observation, averaging inference, and unreliable uncertainty, providing accurate and timely monitoring of transmission dynamics. It offers insights into the impact of different intervention policies and the effectiveness of mass vaccination.
PLOS COMPUTATIONAL BIOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Shuai Jiang, Kan Li, Richard Yi Da Xu
Summary: This paper proposes a novel matrix factorisation algorithm, MBMF, which allows personalized bounding constraints for large-scale datasets. The algorithm constructs a model by constraining the magnitudes of individual feature vectors and converts the bounded optimisation problem into an unconstrained one. Experimental results demonstrate the superiority of MBMF in terms of accuracy and time complexity.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2022)
Article
Computer Science, Hardware & Architecture
Ying Li, Yi Huang, Suranga Seneviratne, Kanchana Thilakarathna, Adriel Cheng, Guillaume Jourjon, Darren Webb, David B. Smith, Richard Yi Da Xu
Summary: This paper explores the potential of fingerprinting internet traffic by sniffing WiFi frames using deep learning methods. The study shows that a hierarchical approach can be used to build a generic traffic classifier that can identify various traffic types and reveal detailed information. Additionally, the use of Multi-Layer Perceptron and Recurrent Neural Networks can identify streaming traffic from encrypted WiFi traffic.
Article
Computer Science, Artificial Intelligence
Yi Huang, Ying Li, Timothy Heyes, Guillaume Jourjon, Adriel Cheng, Suranga Seneviratne, Kanchana Thilakarathna, Darren Webb, Richard Yi Da Xu
Summary: This research proposes a task adaptive siamese neural network for open-set recognition of encrypted network traffic. It introduces generated positive and negative pairs, utilizes Dirichlet Process Gaussian Mixture Model distribution to fit the similarity scores of negative pairs, and constructs a hierarchical cross entropy loss to improve the confidence of the similarity score.
PATTERN RECOGNITION LETTERS
(2022)
Article
Chemistry, Analytical
Andre Pearce, J. Andrew Zhang, Richard Xu
Summary: This paper presents a framework for training a mmWave radar with a camera for labeling and supervising the data. Experimental results demonstrate that the proposed framework consistently achieves high classification accuracy in various environments. The research provides a foundation for future research in unified tracking and sensing systems.
Article
Computer Science, Information Systems
Zhenguo Shi, Qingqing Cheng, J. Andrew Zhang, Richard Yi Da Xu
Summary: This article proposes an innovative scheme, called AFEE-MatNet, for channel state information (CSI)-based human activity recognition (HAR) using deep learning. AFEE-MatNet combines an activity-related feature extraction and enhancement method with a matching network to achieve transferable features and improve recognition performance. The scheme can be directly applied in new/unseen environments without retraining and outperforms existing state-of-the-art HAR methods in terms of accuracy and training time.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Operations Research & Management Science
Qing Yin, Linda Zhong, Yunya Song, Liang Bai, Zhihua Wang, Chen Li, Yida Xu, Xian Yang
Summary: Precision medicine aims to provide personalized healthcare for patients by stratifying them into subgroups based on their health conditions, enabling the development of tailored medical management. In this paper, a deep learning model called ConMEHR is developed for patient stratification from multimodal EHRs. The model effectively aligns and unifies heterogeneous information and outperforms other baseline methods.
ANNALS OF OPERATIONS RESEARCH
(2023)
Article
Business, Finance
Steven Y. K. Wong, Jennifer S. K. Chan, Lamiae Azizi, Richard Y. D. Xu
Summary: This research focuses on the problem of neural network training in a time-varying context. The proposed online early stopping algorithm is shown to outperform current approaches in predicting monthly US stock returns. The study also finds that the predictive power of industry indicators on stock returns varies over time.
INTELLIGENT SYSTEMS IN ACCOUNTING FINANCE & MANAGEMENT
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Ziyue Zhang, Shuai Jiang, Congzhentao Huang, Richard Yi Da Xu
Summary: The paper introduces a novel two-stream network for person Re-identification, which includes a lightweight resolution association module and a self-weighted attention module to handle feature matching under different resolutions. Extensive experiments on multiple datasets demonstrate the effectiveness and superiority of the proposed method.
2021 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP)
(2021)
Proceedings Paper
Computer Science, Interdisciplinary Applications
Steven Y. K. Wong, Jennifer S. K. Chan, Lamiae Azizi, Richard Y. D. Xu
Summary: The paper introduced a method that regularizes the Temporal Convolutional Network using a supervised autoencoder, named the Supervised Temporal Autoencoder (STAE). This approach is beneficial for stock return time-series forecasting and can directly learn features from transformed price series, reducing the need for handcrafted features. The autoencoder also enhances interpretability by allowing users to observe the decoder's output and inspect features retained by the network.
2021 IEEE 45TH ANNUAL COMPUTERS, SOFTWARE, AND APPLICATIONS CONFERENCE (COMPSAC 2021)
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Christos Markos, James J. Q. Yu, Richard Yi Da Xu
Summary: The paper proposes a Bayesian deep learning framework for unsupervised GPS trajectory segmentation, which preprocesses motion feature sequences and utilizes a temporal convolutional neural network for transportation mode identification, addressing issues in intelligent transportation management.
THIRTY-FIFTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, THIRTY-THIRD CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE AND THE ELEVENTH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE
(2021)
Article
Computer Science, Artificial Intelligence
Rui Lv, Dingheng Wang, Jiangbin Zheng, Zhao-Xu Yang
Summary: In this paper, the authors investigate tensor decomposition for neural network compression. They analyze the convergence and precision of tensor mapping theory, validate the rationality of tensor mapping and its superiority over traditional tensor approximation based on the Lottery Ticket Hypothesis. They propose an efficient method called 3D-KCPNet to compress 3D convolutional neural networks using the Kronecker canonical polyadic (KCP) tensor decomposition. Experimental results show that 3D-KCPNet achieves higher accuracy compared to the original baseline model and the corresponding tensor approximation model.
Article
Computer Science, Artificial Intelligence
Xiangkun He, Zhongxu Hu, Haohan Yang, Chen Lv
Summary: In this paper, a novel constrained multi-objective reinforcement learning algorithm is proposed for personalized end-to-end robotic control with continuous actions. The approach trains a single model using constraint design and a comprehensive index to achieve optimal policies based on user-specified preferences.
Article
Computer Science, Artificial Intelligence
Zhijian Zhuo, Bilian Chen, Shenbao Yu, Langcai Cao
Summary: In this paper, a novel method called Expansion with Contraction Method for Overlapping Community Detection (ECOCD) is proposed, which utilizes non-negative matrix factorization to obtain disjoint communities and applies expansion and contraction processes to adjust the degree of overlap. ECOCD is applicable to various networks with different properties and achieves high-quality overlapping community detection.
Article
Computer Science, Artificial Intelligence
Yizhe Zhu, Chunhui Zhang, Jialin Gao, Xin Sun, Zihan Rui, Xi Zhou
Summary: In this work, the authors propose a Contrastive Spatio-Temporal Distilling (CSTD) approach to improve the detection of high-compressed deepfake videos. The approach leverages spatial-frequency cues and temporal-contrastive alignment to fully exploit spatiotemporal inconsistency information.
Review
Computer Science, Artificial Intelligence
Laijin Meng, Xinghao Jiang, Tanfeng Sun
Summary: This paper provides a review of coverless steganographic algorithms, including the development process, known contributions, and general issues in image and video algorithms. It also discusses the security of coverless steganography from theoretical analysis to actual investigation for the first time.
Article
Computer Science, Artificial Intelligence
Yajie Bao, Tianwei Xing, Xun Chen
Summary: Visual question answering requires processing multi-modal information and effective reasoning. Neural-symbolic learning is a promising method, but current approaches lack uncertainty handling and can only provide a single answer. To address this, we propose a confidence based neural-symbolic approach that evaluates NN inferences and conducts reasoning based on confidence.
Article
Computer Science, Artificial Intelligence
Anh H. Vo, Bao T. Nguyen
Summary: Interior style classification is an interesting problem with potential applications in both commercial and academic domains. This project proposes a method named ISC-DeIT, which combines data-efficient image transformer architectures and knowledge distillation, to address the interior style classification problem. Experimental results demonstrate a significant improvement in predictive accuracy compared to other state-of-the-art methods.
Article
Computer Science, Artificial Intelligence
Shashank Kotyan, Danilo Vasconcellos Vargas
Summary: This article introduces a novel augmentation technique called Dynamic Scanning Augmentation to improve the accuracy and robustness of Vision Transformer (ViT). The technique leverages dynamic input sequences to adaptively focus on different patches, resulting in significant changes in ViT's attention mechanism. Experimental results demonstrate that Dynamic Scanning Augmentation outperforms ViT in terms of both robustness to adversarial attacks and accuracy against natural images.
Article
Computer Science, Artificial Intelligence
Hiba Alqasir, Damien Muselet, Christophe Ducottet
Summary: The article proposes a solution to improve the learning process of a classification network by providing shape priors, reducing the need for annotated data. The solution is tested on cross-domain digit classification tasks and a video surveillance application.
Article
Computer Science, Artificial Intelligence
Dexiu Ma, Mei Liu, Mingsheng Shang
Summary: This paper proposes a method using neural dynamics solvers to solve infinity-norm optimization problems. Two improved solvers are constructed and their effectiveness and superiority are demonstrated through theoretical analysis and simulation experiments.
Article
Computer Science, Artificial Intelligence
Francesco Gregoretti, Giovanni Pezzulo, Domenico Maisto
Summary: Active Inference is a computational framework that uses probabilistic inference and variational free energy minimization to describe perception, planning, and action. cpp-AIF is a header-only C++ library that provides a powerful tool for implementing Active Inference for Partially Observable Markov Decision Processes through multi-core computing. It is cross-platform and improves performance, memory management, and usability compared to existing software.
Article
Computer Science, Artificial Intelligence
Zelin Ying, Dawei Cheng, Cen Chen, Xiang Li, Peng Zhu, Yifeng Luo, Yuqi Liang
Summary: This paper proposes a novel stock market trends prediction framework called SMART, which includes a self-supervised stock technical data sequence embedding model S3E. By training with multiple self-supervised auxiliary tasks, the model encodes stock technical data sequences into embeddings and uses the learned sequence embeddings for predicting stock market trends. Extensive experiments on China A-Shares market and NASDAQ market prove the high effectiveness of our model in stock market trends prediction, and its effectiveness is further validated in real-world applications in a leading financial service provider in China.
Article
Computer Science, Artificial Intelligence
Hao Li, Hao Jiang, Dongsheng Ye, Qiang Wang, Liang Du, Yuanyuan Zeng, Liu Yuan, Yingxue Wang, C. Chen
Summary: DHGAT1, a dynamic hyperbolic graph attention network, utilizes hyperbolic metric properties to embed dynamic graphs. It employs a spatiotemporal self-attention mechanism and weighted node representations, resulting in excellent performance in link prediction tasks.
Article
Computer Science, Artificial Intelligence
Jiehui Huang, Zhenchao Tang, Xuedong He, Jun Zhou, Defeng Zhou, Calvin Yu-Chian Chen
Summary: This study proposes a progressive learning multi-scale feature blending model for image deraining tasks. The model utilizes detail dilation and texture extraction to improve the restoration of rainy images. Experimental results show that the model achieves near state-of-the-art performance in rain removal tasks and exhibits better rain removal realism.
Article
Computer Science, Artificial Intelligence
Lizhi Liu, Zilin Gao, Yinhe Wang, Yongfu Li
Summary: This paper proposes a novel discrete-time interconnected model for depicting complex dynamical networks. The model consists of nodes and edges subsystems, which consider the dynamic characteristic of both nodes and edges. By designing control strategies and coupling modes, the stabilization and synchronization of the network are achieved. Simulation results demonstrate the effectiveness of the proposed methods.