Homophily Outlier Detection In Non-Iid Categorical Data,
2021
Singapore Management University
Homophily Outlier Detection In Non-Iid Categorical Data, Guansong Pang, Longbing Cao, Ling Chen
Research Collection School Of Computing and Information Systems
Most of existing outlier detection methods assume that the outlier factors (i.e., outlierness scoring measures) of data entities (e.g., feature values and data objects) are Independent and Identically Distributed (IID). This assumption does not hold in real-world applications where the outlierness of different entities is dependent on each other and/or taken from different probability distributions (non-IID). This may lead to the failure of detecting important outliers that are too subtle to be identified without considering the non-IID nature. The issue is even intensified in more challenging contexts, e.g., high-dimensional data with many noisy features. This work introduces a novel outlier …
Spectral Tensor Train Parameterization Of Deep Learning Layers,
2021
Singapore Management University
Spectral Tensor Train Parameterization Of Deep Learning Layers, A. Obukhov, M. Rakhuba, A. Liniger, Zhiwu Huang, S. Georgoulis, D. Dai, Van Gool L.
Research Collection School Of Computing and Information Systems
We study low-rank parameterizations of weight matrices with embedded spectral properties in the Deep Learning context. The low-rank property leads to parameter efficiency and permits taking computational shortcuts when computing mappings. Spectral properties are often subject to constraints in optimization problems, leading to better models and stability of optimization. We start by looking at the compact SVD parameterization of weight matrices and identifying redundancy sources in the parameterization. We further apply the Tensor Train (TT) decomposition to the compact SVD components, and propose a non-redundant differentiable parameterization of fixed TT-rank tensor manifolds, termed the Spectral Tensor Train Parameterization (STTP). We …
Mixed Dish Recognition With Contextual Relation And Domain Alignment,
2021
Singapore Management University
Mixed Dish Recognition With Contextual Relation And Domain Alignment, Lixi Deng, Jingjing Chen, Chong-Wah Ngo, Qianru Sun, Sheng Tang, Yongdong Zhang, Tat-Seng Chua
Research Collection School Of Computing and Information Systems
Mixed dish is a food category that contains different dishes mixed in one plate, and is popular in Eastern and Southeast Asia. Recognizing the individual dishes in a mixed dish image is important for health related applications, e.g. to calculate the nutrition values of the dish. However, most existing methods that focus on single dish classification are not applicable to the recognition of mixed dish images. The main challenge of mixed dish recognition comes from three aspects: a wide range of dish types, the complex dish combination with severe overlap between different dishes and the large visual variances of same …
Efficient Retrieval Of Matrix Factorization-Based Top-K Recommendations: A Survey Of Recent Approaches,
2021
Singapore Management University
Efficient Retrieval Of Matrix Factorization-Based Top-K Recommendations: A Survey Of Recent Approaches, Dung D. Le, Hady W. Lauw
Research Collection School Of Computing and Information Systems
Top-k recommendation seeks to deliver a personalized list of k items to each individual user. An established methodology in the literature based on matrix factorization (MF), which usually represents users and items as vectors in low-dimensional space, is an effective approach to recommender systems, thanks to its superior performance in terms of recommendation quality and scalability. A typical matrix factorization recommender system has two main phases: preference elicitation and recommendation retrieval. The former analyzes user-generated data to learn user preferences and item characteristics in the form of latent feature vectors, whereas the latter ranks the candidate items based on the …
Powered By Ai,
2021
Michigan Dental Association
Powered By Ai, Christopher J. Smiley
The Journal of the Michigan Dental Association
Artificial Intelligence (AI) is revolutionizing dental practice through its ability to process vast amounts of data, enhance diagnosis, and improve patient care. However, AI introduces the challenge of bias and ethical considerations. Dentists and dental benefit providers are utilizing AI for early disease detection and efficient data management, but transparency and fairness in AI algorithms are vital. The Rome Call for AI Ethics emphasizes ethical, non-biased AI development. In the broader context, AI-driven marketing and predictive behavior raise concerns about privacy and ethical data use. The dental community must embrace AI's power while upholding ethical standards and transparency.
Time Period-Based Top-K Semantic Trajectory Pattern Query,
2021
Singapore Management University
Time Period-Based Top-K Semantic Trajectory Pattern Query, Munkh-Erdene Yadamjav, Farhana Murtaza Choudhury, Zhifeng Bao, Baihua Zheng
Research Collection School Of Computing and Information Systems
The sequences of user check-ins form semantic trajectories that represent the movement of users through time, along with the types of POIs visited. Extracting patterns in semantic trajectories can be widely used in applications such as route planning and trip recommendation. Existing studies focus on the entire time duration of the data, which may miss some temporally significant patterns. In addition, they require thresholds to define the interestingness of the patterns. Motivated by the above, we study a new problem of finding top-k semantic trajectory patterns w.r.t. a given time period and categories by considering the spatial closeness of POIs. …
Cross-Topic Rumor Detection Using Topic-Mixtures,
2021
Singapore Management University
Cross-Topic Rumor Detection Using Topic-Mixtures, Weijieying Ren, Jing Jiang, Ling Min Serena Khoo, Hai Leong Chieu
Research Collection School Of Computing and Information Systems
There has been much interest in rumor detection using deep learning models in recent years. A well-known limitation of deep learning models is that they tend to learn superficial patterns, which restricts their generalization ability. We find that this is also true for cross-topic rumor detection. In this paper, we propose a method inspired by the “mixture of experts” paradigm. We assume that the prediction of the rumor class label given an instance is dependent on the topic distribution of the instance. After deriving a vector representation for each topic, given an instance, we derive a “topic mixture” vector for …
Variable Autoencoders For Biosensor Data Augmentation,
2021
Andrews University
Variable Autoencoders For Biosensor Data Augmentation, Solomon Kim
Honors Theses
Over the past decade machine learning and artificial intelligence's resurgence spawned the desire to mimic human creative ability. Initially attempts to create images, music, and text flooded the community, though little has been learned regarding constrained, one-dimensional data generation. This paper demonstrates a variational autoencoder approach to this problem. By modeling biosensor current and concentration data we aim to augment the existing dataset. In training a multi-layer neural network based encoder and decoder we were able to generate realistic, original samples., These results demonstrate the ability to realistically augment datasets, improving training of machine learning models designed to predict concentration …
Learning To Fuse Asymmetric Feature Maps In Siamese Trackers,
2021
Beijing Institute of Technology
Learning To Fuse Asymmetric Feature Maps In Siamese Trackers, Wencheng Han, Xingping Dong, Fahad Shahbaz Khan, Ling Shao, Jianbing Shen
Computer Vision Faculty Publications
Recently, Siamese-based trackers have achieved promising performance in visual tracking. Most recent Siamese-based trackers typically employ a depth-wise cross-correlation (DW-XCorr) to obtain multi-channel correlation information from the two feature maps (target and search region). However, DW-XCorr has several limitations within Siamese-based tracking: it can easily be fooled by distractors, has fewer activated channels and provides weak discrimination of object boundaries. Further, DW-XCorr is a handcrafted parameter-free module and cannot fully benefit from offline learning on large-scale data. We propose a learnable module, called the asymmetric convolution (ACM), which learns to better capture the semantic correlation information in offline training on …
Deep Gaussian Processes For Few-Shot Segmentation,
2021
Linköping University & Zenseact Ab
Deep Gaussian Processes For Few-Shot Segmentation, Joakim Johnander, Johan Edstedt, Martin Danelljan, Michael Felsberg, Fahad Shahbaz Khan
Computer Vision Faculty Publications
Few-shot segmentation is a challenging task, requiring the extraction of a generalizable representation from only a few annotated samples, in order to segment novel query images. A common approach is to model each class with a single prototype. While conceptually simple, these methods suffer when the target appearance distribution is multi-modal or not linearly separable in feature space. To tackle this issue, we propose a few-shot learner formulation based on Gaussian process (GP) regression. Through the expressivity of the GP, our approach is capable of modeling complex appearance distributions in the deep feature space. The GP provides a principled way …
On Generating Transferable Targeted Perturbations,
2021
Australian National University
On Generating Transferable Targeted Perturbations, Muzammal Naseer, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, Fatih Porikli
Computer Vision Faculty Publications
While the untargeted black-box transferability of adversarial perturbations has been extensively studied before, changing an unseen model's decisions to a specific 'targeted' class remains a challenging feat. In this paper, we propose a new generative approach for highly transferable targeted perturbations (TTP). We note that the existing methods are less suitable for this task due to their reliance on class-boundary information that changes from one model to another, thus reducing transferability. In contrast, our approach matches the perturbed image 'distribution' with that of the target class, leading to high targeted transferability rates. To this end, we propose a new objective …
Orthogonal Projection Loss,
2021
Mohamed bin Zayed University of Artificial Intelligence
Orthogonal Projection Loss, Kanchana Ranasinghe, Muzammal Naseer, Munawar Hayat, Salman Khan, Fahad Shahbaz Khan
Computer Vision Faculty Publications
Deep neural networks have achieved remarkable performance on a range of classification tasks, with softmax cross-entropy (CE) loss emerging as the de-facto objective function. The CE loss encourages features of a class to have a higher projection score on the true class-vector compared to the negative classes. However, this is a relative constraint and does not explicitly force different class features to be well-separated. Motivated by the observation that ground-truth class representations in CE loss are orthogonal (one-hot encoded vectors), we develop a novel loss function termed 'Orthogonal Projection Loss' (OPL) which imposes orthogonality in the feature space. OPL augments …
Survey On Quantum Circuit Compilation For Noisy Intermediate-Scale Quantum Computers: Artificial Intelligence To Heuristics,
2021
CUNY New York City College of Technology
Survey On Quantum Circuit Compilation For Noisy Intermediate-Scale Quantum Computers: Artificial Intelligence To Heuristics, Janusz Kusyk, Samah Mohamed Saeed, Muharrem Umit Uyar
Publications and Research
Computationally expensive applications, including machine learning, chemical simulations, and financial modeling, are promising candidates for noisy intermediate scale quantum (NISQ) computers. In these problems, one important challenge is mapping a quantum circuit onto NISQ hardware while satisfying physical constraints of an underlying quantum architecture. Quantum circuit compilation (QCC) aims to generate feasible mappings such that a quantum circuit can be executed in a given hardware platform with acceptable confidence in outcomes. Physical constraints of a NISQ computer change frequently, requiring QCC process to be repeated often. When a circuit cannot directly be executed on a quantum hardware due to its …
Frequency Coordinated Control Strategy Of Microgrid Based On Fuzzy Prediction,
2021
Xuchang Electrical Vocational College, Xuchang 461000, China;
Frequency Coordinated Control Strategy Of Microgrid Based On Fuzzy Prediction, Kunping Zhang, Hao Lin
Journal of System Simulation
Abstract: Aiming at the problem of frequency fluctuation of wind power generation connected to microgrid, a frequency coordinated control strategy based on model predictive control (MPC) is proposed. In this strategy, the wind turbine and plug-in hybrid electric vehicle (PHEV) are included in the frequency control system. The pitch angle of the fan and the charge and discharge of PHEV are controlled to adjust the grid frequency and supplement the frequency modulation resources of microgrid. WTG pitch angle control system and PHEV power control system are modeled, and their control principles are described. In order to prevent excessive use of …
Guaranteed Cost Preview And Repetitive Control For Uncertain Linear Discrete Time-Delay Systems,
2021
School of information engineering, Xiangtan University, Xiangtan 411105, China;
Guaranteed Cost Preview And Repetitive Control For Uncertain Linear Discrete Time-Delay Systems, Yonghong Lan, Jinlin He
Journal of System Simulation
Abstract: For a class of uncertain linear discrete time-delay systems, a design method for guaranteed cost preview and repetitive controller is proposed . By introducing a repetitive controller in the forward channel to improve the tracking accuracy of the system, L-order difference operators are used to construct an augmented error system that contains preview information but does not include time delay, and the design problem of guaranteed cost preview and repetitive controller is converted into an output feedback adjustment problem. Using the Lyapunov stability theory and the linear matrix inequality method, the sufficient conditions for guaranteeing the asymptotic stability of …
Neural Network Optimized Sensorless Permanent Magnet Synchronous Motor Control System,
2021
School of mechanical engineering, University of Shanghai for Science and Technology, Shanghai 200093, China;
Neural Network Optimized Sensorless Permanent Magnet Synchronous Motor Control System, Lixin Ma, Yongjie Zhu, Leyan Ji
Journal of System Simulation
Abstract: In order to solve the poor accuracy of the speed and rotor position of permanent magnet synchronous motor caused by sensor, a sensorless control system is proposed to calculate the speed and rotor position of PMSM with extended Kalman filtering algorithm. BP neural network algorithm is used to optimize the covariance matrix Q and R of EKF, which improves the accurate calculation values of rotational speed and rotor position. At the same time, the speed sliding mode controller combined with the current feed-forward decoupling unit are used to improve the stability of the whole control system. The simulation results …
An Improved Social Force Model For Crowd Simulation,
2021
School of information and control engineering, Xi'an University of Architecture and Technology, Xi'an 710055, China;
An Improved Social Force Model For Crowd Simulation, Changhua Li, Yang Jing, Zhijie Li
Journal of System Simulation
Abstract: In view of the traditional social force model, it is difficult to deal with the problems of single pedestrian trajectory and loose crowd in the process of crowd evacuation, and an improved social force model is proposed. Based on the original social force model, the movement trajectory of the person is changed by considering the choice of the movement direction The intensity of panic and attraction in the process of pedestrian evacuation is considered to reproduce the self-organizing behavior in the process of pedestrian evacuation, and the simulations are performed in individual and group mode. The authenticity of the …
Research On The Method Of Operational Concept Description Based On Sysml,
2021
Beijing Institute of System Engineering, Beijing 100101, China;
Research On The Method Of Operational Concept Description Based On Sysml, Siming Peng, Xiao Gang, Qingzhang Yu, Zeming Li
Journal of System Simulation
Abstract: For the convenience of understanding and communication between researchers among different domains, the standardized method for operational concept description is preferred. Hence, based on the principles of systems architecture, the System Model Language (SysML) is proposed for the visualized and standardized description of operational concept. The form of combination for Department of Defense Architecture Framework (DoDAF) and SysML during the description of operational concept is analyzed, and the multi-view point products are used to descript the operational background, capability requirement and systems architecture as well as the operational activity of operational concept. The "Distributed Lethality" is utilized as an …
Simulation Platform For Source-Load Control Of Active Power Based On Modular Architecture,
2021
1. North China Electric Power University, State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources, Beijing 102206, China; ;
Simulation Platform For Source-Load Control Of Active Power Based On Modular Architecture, Hu Yang, Wenying Liu, Liping Zhu, Li Xiao, Weizhou Wang
Journal of System Simulation
Abstract: The integrated proportion of wind power is increasing year by year, and the source-load coordinated control of active power can effectively improve the level of wind power consumption. In order to ensure the effective application of the strategy, a simulation platform based on modular architecture is developed for source-load control of active power, including SQL Server database, timing control module of data interaction, calculation module of source-load control strategy, and output display module. The simulation platform solves the automatic control of data interaction timing in the process of source-load control, and visualizes the effect of the source-load control, so …
Path Designing Of Multi-Omnidirectional Wheel Collaborative Sorting Platform,
2021
1. School of Electronic Information and Artificial Intelligence, Shaanxi University of Science and Technology, Xi'an 710021, China; ;
Path Designing Of Multi-Omnidirectional Wheel Collaborative Sorting Platform, Li Qi, Wang Wei
Journal of System Simulation
Abstract: Aiming at the problems of low efficiency, high labor cost and low flexibility of traditional logistics sorting system, an automatic logistics sorting system is designed. The improved A* algorithm and the artificial potential field method are used to realize the automatic path planning of the system by taking the transportation path as the research object. The A* algorithm is improved by adjusting the weights of actual cost and estimated cost, and the artificial potential field method is improved by adding virtual sub-target points and adjusting adaptive parameters, so as to complete the function of path planning of goods. Simulation …
