Artificial Intelligence and Robotics Commons^™

Notmad: Estimating Bayesian Networks With Sample-Specific Structures And Parameters, Benjamin Lengerich, Caleb Ellington, Bryon Aragam, Eric P. Xing, Manolis Kellis

Machine Learning Faculty Publications

Context-specific Bayesian networks (i.e. directed acyclic graphs, DAGs) identify context-dependent relationships between variables, but the non-convexity induced by the acyclicity requirement makes it difficult to share information between context-specific estimators (e.g. with graph generator functions). For this reason, existing methods for inferring context-specific Bayesian networks have favored breaking datasets into subsamples, limiting statistical power and resolution, and preventing the use of multidimensional and latent contexts. To overcome this challenge, we propose NOTEARS-optimized Mixtures of Archetypal DAGs (NOTMAD). NOTMAD models context-specific Bayesian networks as the output of a function which learns to mix archetypal networks according to sample context. The archetypal …

Protection Of Patient Privacy On Mobile Device Machine Learning, Matthew Nguyen

Cybersecurity Undergraduate Research Showcase

An existing StudentLife Study mobile dataset was evaluated and organized to be applied to different machine learning methods. Different variables like user activity, exercise, sleep, study space, social, and stress levels are optimized to train a model that could predict user stress level. The different machine learning methods would test if both patient data privacy and training efficiency can be ensured.

Transfer-Learned Pruned Deep Convolutional Neural Networks For Efficient Plant Classification In Resource-Constrained Environments, Martinson Ofori

Masters Theses & Doctoral Dissertations

Traditional means of on-farm weed control mostly rely on manual labor. This process is time-consuming, costly, and contributes to major yield losses. Further, the conventional application of chemical weed control can be economically and environmentally inefficient. Site-specific weed management (SSWM) counteracts this by reducing the amount of chemical application with localized spraying of weed species. To solve this using computer vision, precision agriculture researchers have used remote sensing weed maps, but this has been largely ineffective for early season weed control due to problems such as solar reflectance and cloud cover in satellite imagery. With the current advances in artificial …

Information Extraction And Classification On Journal Papers, Lei Yu

School of Computing: Dissertations, Theses, and Student Research

The importance of journals for diffusing the results of scientific research has increased considerably. In the digital era, Portable Document Format (PDF) became the established format of electronic journal articles. This structured form, combined with a regular and wide dissemination, spread scientific advancements easily and quickly. However, the rapidly increasing numbers of published scientific articles requires more time and effort on systematic literature reviews, searches and screens. The comprehension and extraction of useful information from the digital documents is also a challenging task, due to the complex structure of PDF.

To help a soil science team from the United States …

Development Of Sensor, Sensory System And Signal Processing Algorithm For Intelligent Sensing Applications, Xingzhe Zhang

Dissertations

Sensors have been receiving significant attention in the last decade and the demand for sensory systems has increased in recent years due to the rapid growth in the field of artificial intelligence (AI). Sensors can improve people’s awareness by providing them with real-time information on the environment and their immediate health conditions. This dissertation presents the fulfilment of three main projects and focuses on the development of a sensor, a sensory system, and a sensor signal recognition system for AI applications by employing printed electronics, analog circuit design, and digital signal processing techniques.

In the first project, a multi-channel stethograph …

Generating Music With Sentiments, Chunhui Bao

Dissertations and Theses Collection (Open Access)

In this thesis, I focus on the music generation conditional on human sentiments such as positive and negative. As there are no existing large-scale music datasets annotated with sentiment labels, generating high-quality music conditioned on sentiments is hard. I thus build a new dataset consisting of the triplets of lyric, melody and sentiment, without requiring any manual annotations. I utilize an automated sentiment recognition model (based on the BERT trained on Edmonds Dance dataset) to "label'' the music according to the sentiments recognized from its lyrics. I then train the model of generating sentimental music and call the method Sentimental …

Artificial Intelligence As Augmenting Automation: Implications For Employment, F. Ted Tschang, Esteve Almirall

Research Collection Lee Kong Chian School Of Business

There has been great concern in recent years that artificial intelligence (AI) may cause widespread unemployment, but proponents say that AI augments existing jobs. Both of these positions have substance, but there is a need is to articulate the mechanisms by which AI may actually do both, and in the process, transform work and business organizations alike. We use economic studies showing past transformations automation wrought on the structure of employment and skills (such as the favouring of nonroutine skills) to articulate a ground for discussion. We then use case evidence of AI and automation to show how AI is …

Learning Knowledge-Enriched Company Embeddings For Investment Management, Gary Ang, Ee-Peng Lim

Research Collection School Of Computing and Information Systems

Relationships between companies serve as key channels through which the effects of past stock price movements and news events propagate and influence future price movements. Such relationships can be implicitly found in knowledge bases or explicitly represented as knowledge graphs. In this paper, we propose KnowledgeEnriched Company Embedding (KECE), a novel multi-stage attentionbased dynamic network embedding model combining multimodal information of companies with knowledge from Wikipedia and knowledge graph relationships from Wikidata to generate company entity embeddings that can be applied to a variety of downstream investment management tasks. Experiments on an extensive set of real-world stock prices and news …

Span-Level Emotion Cause Analysis With Neural Sequence Tagging, Xiangju Li, Wei Gao, Shi Feng, Daling Wang, Shafiq Joty

Research Collection School Of Computing and Information Systems

This paper addresses the task of span-level emotion cause analysis (SECA). It is a finer-grained emotion cause analysis (ECA) task, which aims to identify the specific emotion cause span(s) behind certain emotions in text. In this paper, we formalize SECA as a sequence tagging task for which several variants of neural network-based sequence tagging models to extract specific emotion cause span(s) in the given context. These models combine different types of encoding and decoding approaches. Furthermore, to make our models more "emotionally sensitive'', we utilize the multi-head attention mechanism to enhance the representation of context. Experimental evaluations conducted on two …

Fleet Sizing And Allocation For On-Demand Last-Mile Transportation Systems, Karmel Shehadeh, Hai Wang, Peter Zhang

Research Collection School Of Computing and Information Systems

The last-mile problem refers to the provision of travel service from the nearest public transportation node to home or other destination. Last-Mile Transportation Systems (LMTS), which have recently emerged, provide on-demand shared transportation. In this paper, we investigate the fleet sizing and allocation problem for the on-demand LMTS. Specifically, we consider the perspective of a last-mile service provider who wants to determine the number of servicing vehicles to allocate to multiple last-mile service regions in a particular city. In each service region, passengers demanding last-mile services arrive in batches, and allocated vehicles deliver passengers to their final destinations. The passenger …

Predicting Anti-Asian Hateful Users On Twitter During Covid-19, Jisun An, Haewoon Kwak, Claire Seungeun Lee, Bogang Jun, Yong-Yeol Ahn

Research Collection School Of Computing and Information Systems

We investigate predictors of anti-Asian hate among Twitter users throughout COVID-19. With the rise of xenophobia and polarization that has accompanied widespread social media usage in many nations, online hate has become a major social issue, attracting many researchers. Here, we apply natural language processing techniques to characterize social media users who began to post anti-Asian hate messages during COVID-19. We compare two user groups—those who posted anti-Asian slurs and those who did not—with respect to a rich set of features measured with data prior to COVID-19 and show that it is possible to predict who later publicly posted anti-Asian …

Stock Market Trend Forecasting Based On Multiple Textual Features: A Deep Learning Method, Zhenda Hu, Zhaoxia Wang, Seng-Beng Ho, Ah-Hwee Tan

Research Collection School Of Computing and Information Systems

Stock market trend forecasting is a valuable and challenging research task for both industry and academia. In order to explore the influence of stock news information on the stock market trend, a textual embedding construction method is proposed to encode multiple textual features, including topic features, sentiment features, and semantic features extracted from stock news textual content. In addition, a deep learning method is designed by using financial data and multiple textual features obtained from multiple news textual embeddings for short-term stock market trend prediction. For evaluation, extensive experiments on real stock market data are conducted. The experimental results illustrate …

Self-Supervised Multi-Class Pre-Training For Unsupervised Anomaly Detection And Segmentation In Medical Images, Yu Tian, Fengbei Liu, Guansong Pang, Yuanhong Chen, Yuyuan Liu, Johan W. Verjans, Rajvinder Singh

Research Collection School Of Computing and Information Systems

Unsupervised anomaly detection (UAD) that requires only normal (healthy) training images is an important tool for enabling the development of medical image analysis (MIA) applications, such as disease screening, since it is often difficult to collect and annotate abnormal (or disease) images in MIA. However, heavily relying on the normal images may cause the model training to overfit the normal class. Self-supervised pre-training is an effective solution to this problem. Unfortunately, current self-supervision methods adapted from computer vision are sub-optimal for MIA applications because they do not explore MIA domain knowledge for designing the pretext tasks or the training process. …

Towards Enriching Responses With Crowd-Sourced Knowledge For Task-Oriented Dialogue, Yingxu He, Lizi Liao, Zheng Zhang, Tat-Seng Chua

Research Collection School Of Computing and Information Systems

Task-oriented dialogue agents are built to assist users in completing various tasks. Generating appropriate responses for satisfactory task completion is the ultimate goal. Hence, as a convenient and straightforward way, metrics such as success rate, inform rate etc., have been widely leveraged to evaluate the generated responses. However, beyond task completion, there are several other factors that largely affect user satisfaction, which remain under-explored. In this work, we focus on analyzing different agent behavior patterns that lead to higher user satisfaction scores. Based on the findings, we design a neural response generation model EnRG. It naturally combines the power of …

The Ratio Method: Addressing Complex Tort Liability In The Fourth Industrial Revolution, Harrison C. Margolin, Grant H. Frazier

St. Mary's Law Journal

Emerging technologies of the Fourth Industrial Revolution show fundamental promise for improving productivity and quality of life, though their misuse may also cause significant social disruption. For example, while artificial intelligence will be used to accelerate society’s processes, it may also displace millions of workers and arm cybercriminals with increasingly powerful hacking capabilities. Similarly, human gene editing shows promise for curing numerous diseases, but also raises significant concerns about adverse health consequences related to the corruption of human and pathogenic genomes.

In most instances, only specialists understand the growing intricacies of these novel technologies. As the complexity and speed of …

Generating Synthetic Training Data For Deep Learning-Based Uav Trajectory Prediction, Brendan T. Morris, Stefan Becker, Ronny Hug, Wolfgang Huebner, Michael Arens

Electrical & Computer Engineering Faculty Research

Deep learning-based models, such as recurrent neural networks (RNNs), have been applied to various sequence learning tasks with great success. Following this, these models are increasingly replacing classic approaches in object tracking applications for motion prediction. On the one hand, these models can capture complex object dynamics with less modeling required, but on the other hand, they depend on a large amount of training data for parameter tuning. Towards this end, we present an approach for generating synthetic trajectory data of unmanned-aerial-vehicles (UAVs) in image space. Since UAVs, or rather quadrotors are dynamical systems, they can not follow arbitrary trajectories. …

Research On Cgf-Oriented Intention Recognition Behavioral Modeling Framework, Xu Kai, Yunxiu Zeng, Wansen Wu, Quanjun Yin, Yabing Zha

Journal of System Simulation

Abstract: As an important cognitive behavior in Computer Generated Forces (CGF), Intention Recognition reasons the temporal relations between actions of friends and enemies to recognize their true intentions, and provides the observer with far more focused decision-making ability. In order to further formalize the modeling of CGF-oriented intention recognition, the paper reviews the worldwide research development from 1980s, along with the designs and implementations of different methods. Following the theory of Situation Awareness, the paper analyzes the situation awareness process of CGF, its impacting factors and constraints and proposes a generalized intention recognition framework considering different problem characteristics, constraints and …

Identification Of Main Steam Temperature System Based On Improved Particle Swarm Optimization, Zhenqian Cao, Yin Jiang, Jinhua Zhang

Journal of System Simulation

Abstract: Establishing an accurate mathematical model of main steam temperature is the basis of improving the performance of control system. Aiming at the problems of early maturity and slow convergence in traditional particle swarm optimization (PSO) algorithm in model identification, an improved PSO algorithm with shrinkage factor is proposed. The algorithm improves the global optimization capability and convergence speed of the algorithm by adjusting the shrinkage factor. The on-site operating data of a 350 MW circulating fluidized bed (CFB) boiler in a power plant in Shanxi province are used in the identification of the main steam model parameters, and the …

Dqn-Based Path Planning Method And Simulation For Submarine And Warship In Naval Battlefield, Xiaodong Huang, Haitao Yuan, Bi Jing, Liu Tao

Journal of System Simulation

Abstract: To realize multi-agent intelligent planning and target tracking in complex naval battlefield environment, the work focuses on agents (submarine or warship), and proposes a simulation method based on reinforcement learning algorithm called Deep Q Network (DQN). Two neural networks with the same structure and different parameters are designed to update real and predicted Q values for the convergence of value functions. An ε-greedy algorithm is proposed to design an action selection mechanism, and a reward function is designed for the naval battlefield environment to increase the update velocity and generalization ability of Learning with Experience Replay (LER). Simulation results …

Research On Path Tracking Control Strategy Of Four-Wheel Steering Intelligent Vehicle, Jingbo Zhao, Liangpeng Zhu, Chengye Liu

Journal of System Simulation

Abstract: Aiming at the instability of path tracking control of intelligent vehicle at high speed, a path tracking control strategy of four-wheel steering combined with differential braking is proposed. In the upper layer, the front wheel active steering controller is designed based on the path tracking model. In the lower layer, the integrated controller of active rear steering and additional yaw moment is designed using the sliding mode control method. The additional yaw moment is transformed into the control of single wheel by designing differential braking distribution strategy. Simulation results show that the tracking accuracy of the combined control strategy …

Fixed-Time Event-Triggered Formation Control For Multiple Uavs, Ye Shuai, Guoping Jiang, Yingjiang Zhou, Liu Shang

Journal of System Simulation

Abstract: To solve the quadrotor UAV formation control problem, an event-triggered fixed time formation control algorithm of multi quadrotor UAV system is studied. For the attitude loop control problem of UAV, the switching fixed time sliding surface is selected to make the system reach the equilibrium point in a fixed time when the system state is on the sliding surface. A fixed-time sliding mode controller is designed so that the system state which is not on the sliding surface can reach the sliding surface within a fixed time. In view of the position loop control problem of UAV, the event-driven …

Virus Propagation And System Simulation Based On Cellular Automata Model, Lijuan Zhang, Fuchang Wang, Zhengang Li

Journal of System Simulation

Abstract: A susceptible-latent-infected-cured-immune virus spreading model is established according to the characteristics of epidemic transmission and the actual urban spatial map model. Individuals in the environment are regarded as agents, and the spreading mechanism is established according to the principle of cellular. The effects of different strategies and different characteristics of virus spreading on epidemic have been studied. The role and effect of important factors on epidemic prevention and control method are discussed, and the model is validated taking Shijiazhuang epidemic as an example . The results show that the number of initial latent, the infectivity of disease, vaccination proportion …

Research On Accurate Gesture Recognition Algorithm In Complex Environment Based On Machine Vision, Xu Sheng, Wenyu Feng, Zhicheng Liu, Xintao Tu, Minrui Fei, Kun Zhang

Journal of System Simulation

Abstract: To address the issue of cross infection caused by elevator public buttons during COVID-19, a software algorithm based on machine vision for non-contact control of public buttons by gesture recognition is designed. In order to improve the accuracy of gesture recognition, an improved YOLOv4 algorithm is proposed. A Ghost module is designed based on attention mechanism, and the ResBlock module in YOLOv4 is improved to Ghost module. The experimental results show that, in the task of gesture recognition, the detection speed is improved by 14% and the detection accuracy is improved by 0.1% compared with the original model. The …

Simulation Of Civil Aircraft Takeoff Scenario Based On Mbse, Liangyu Zhao, Junjie Ye, He Qi, Guo Wei, Zhao Yong

Journal of System Simulation

Abstract: Aiming at the interactions and flaws of design being difficult to be fully discovered, the requirements being hard to be traced, the early verification of system design being difficult to be realized and so on, a Model-Based System Engineering (MBSE) method is adopted to realize the simulation of civil aircraft take-off scenario. Based on the analysis of civil aircraft takeoff scenario requirements, the civil aircraft takeoff scenario simulation architecture, take-off scenario discrete logic model, and continuous physical simulation model are established. The method of fusing the SysML model and Simulink model and the 3D visualization of simulation data …

An Equivalent Method Of Uav Simulating High-Altitude Reconnaissance Equipment Optical Imaging, Jiangang Tu, Xu Cheng, Wang Hui, Zenghui Shen

Journal of System Simulation

Abstract: Aiming at the situation that the actual image of enemy reconnaissance equipment is difficult to be obtained, and the process of using similar equipment of our army is complicated and expensive in the detection of the camouflage effect of our important military targets, an equivalent detection method of using the UAV aerial photography to generate the high-altitude optical reconnaissance images is proposed. The number of pixels in the equivalent image is determined by comparing the acquisition height and the optical imaging device’s parameters of the UAV and the high-altitude reconnaissance equipment. The compressed pixel value is calculated through the …

Joint Distribution Location-Routing Problem And Large Neighborhood Search Algorithm, Zhenping Li, Yuwei Zhao, Yuwei Zhang, Lining Xing, Ren Teng

Journal of System Simulation

Abstract: Based on the characteristics of two-echelon, multi-center, and heterogeneous fleets in urban logistics joint distribution system, the two-echelon joint distribution location routing problem is studied. The problem is formulated into a mixed integer programming model to minimize the total costs. An adaptive large neighborhood search algorithm (ALNS) for solving the model with multiple deletion and insertion operators is proposed to obtain neighborhood solution. The selection probability of each operator is adjusted according to the neighborhood solution to accelerate the convergence speed. Several test examples are generated based on the benchmark of location-routing problem. Both ALNS algorithm and Gurobi software …

Modeling And Simulation Of Crowd Multi-Perception Behavior Model In Closed Public Places, Ma Jun, Hu Jun

Journal of System Simulation

Abstract: The social forces-based multi-agent modeling method is used to build the crowd simulation models in most of the existing crowd behavior simulation software systems. From the perspective of cognitive psychology, combined with the spatial logic model of the generalized influence of the crowd obtained by the author’s previous research, and based on the sigmoid function, the mathematical model of the individual visual and auditory perception of the crowd is constructed. Taking the typical subway transfer hall as an example, the objective function of large passenger volume evacuation optimization is established to construct the crowd evacuation simulation model. The simulation …

High Performance Simulation Method And Test Platform For Full Envelop Of Flight Vehicle, Jianlin Wang, Sufang Chen, Tie Ming, Liu Jing, Zhang Jun

Journal of System Simulation

Abstract: In order to meet the needs of the overall performance simulation and verification of the vehicle with large subsamples and massive states, a high-performance simulation method is proposed, in which a parallel simulation architecture based on high-performance computers is established, the technologies and software such as decoupling of simulation tasks, real-time monitoring of simulation, concurrent conflict control and automatic aggregation of large-scale test results are realized, and a test platform for the large-scale parallel verification of flight performance is developed. Simulation results show that the method has high acceleration ratio and scalability, and can complete the verification efficiently, …

Research On Simulation And Optimization Of Maintenance Support Mode And Human Resources Auocation For Complex Equipment, Junhai Cao, Yiming Guo, Chuang Zhang, Qingyi Guo

Journal of System Simulation

Abstract: Aiming at the single maintenance support mode of the army’s complex equipment and the strong randomness of human resource allocation, the military-level equipment maintenance process is taken as the research object, the complex equipment maintenance plan, maintenance support mode, human resource configuration, and maintenance process flow are analyzed. The simulation model of a complex equipment maintenance workshop is constructed, the influence of maintenance plans, maintenance support modes, and human resource allocation on indicators of “maintenance support time” and “human resource satisfaction rate” is explored, the internal laws is analyzed and the manpower optimal allocation for the maintenance of …

Survey Of Evolutionary Behavior Tree Algorithm, Yang Jie, Zhang Qi, Junjie Zeng, Quanjun Yin

Journal of System Simulation

Abstract: Evolutionary behavior tree method is an agent behavior modeling method which uses evolutionary algorithm to generate and optimize behavior tree model. Based on the background knowledge of behavior tree and evolutionary algorithm, three kinds of evolutionary behavior tree algorithms based on genetic programming, grammar evolution and hybrid algorithm as well as corresponding improved algorithms are described, and the advantages and disadvantages of different algorithms are analyzed and compared. The specific applications of evolutionary behavior tree in combat simulation, game artificial intelligence, robotics and other fields are summarized. The future development trends of evolutionary behavior tree are proposed and discussed …