Artificial Intelligence and Robotics Commons^™

Distributed Learning With Automated Stepsizes, Benjamin Liggett

All Theses

Stepsizes for optimization problems play a crucial role in algorithm convergence, where the stepsize must undergo tedious manual tuning to obtain near-optimal convergence. Recently, an adaptive method for automating stepsizes was proposed for centralized optimization. However, this method is not directly applicable to decentralized optimization because it allows for heterogeneous agent stepsizes. Furthermore, directly using consensus between agent stepsizes to mitigate stepsize heterogeneity can decrease performance and even lead to divergence.

This thesis proposes an algorithm to remedy the tedious manual tuning of stepsizes in decentralized optimization. Our proposed algorithm automates the stepsize and uses dynamic consensus between agents’ stepsizes …

Development Of Graphical Models And Statistical Physics Motivated Approaches To Genomic Investigations, Yashwanth Lagisetty

Dissertations and Theses (Open Access)

Identifying genes involved in disease pathology has been a goal of genomic research since the early days of the field. However, as technology improves and the body of research grows, we are faced with more questions than answers. Among these is the pressing matter of our incomplete understanding of the genetic underpinnings of complex diseases. Many hypotheses offer explanations as to why direct and independent analyses of variants, as done in genome-wide association studies (GWAS), may not fully elucidate disease genetics. These range from pointing out flaws in statistical testing to invoking the complex dynamics of epigenetic processes. In the …

Unsupervised Contrastive Representation Learning For Knowledge Distillation And Clustering, Fei Ding

All Dissertations

Unsupervised contrastive learning has emerged as an important training strategy to learn representation by pulling positive samples closer and pushing negative samples apart in low-dimensional latent space. Usually, positive samples are the augmented versions of the same input and negative samples are from different inputs. Once the low-dimensional representations are learned, further analysis, such as clustering, and classification can be performed using the representations. Currently, there are two challenges in this framework. First, the empirical studies reveal that even though contrastive learning methods show great progress in representation learning on large model training, they do not work well for small …

Reduced Fuel Emissions Through Connected Vehicles And Truck Platooning, Paul D. Brummitt

Electronic Theses and Dissertations

Vehicle-to-infrastructure (V2I) and vehicle-to-vehicle (V2V) communication enable the sharing, in real time, of vehicular locations and speeds with other vehicles, traffic signals, and traffic control centers. This shared information can help traffic to better traverse intersections, road segments, and congested neighborhoods, thereby reducing travel times, increasing driver safety, generating data for traffic planning, and reducing vehicular pollution. This study, which focuses on vehicular pollution, used an analysis of data from NREL, BTS, and the EPA to determine that the widespread use of V2V-based truck platooning—the convoying of trucks in close proximity to one another so as to reduce air drag …

Using Ensemble Learning Techniques To Solve The Blind Drift Calibration Problem, Devin Scott Drake

Computer Science Theses & Dissertations

Large sets of sensors deployed in nearly every practical environment are prone to drifting out of calibration. This drift can be sensor-based, with one or several sensors falling out of calibration, or system-wide, with changes to the physical system causing sensor-reading issues. Recalibrating sensors in either case can be both time and cost prohibitive. Ideally, some technique could be employed between the sensors and the final reading that recovers the drift-free sensor readings. This paper covers the employment of two ensemble learning techniques — stacking and bootstrap aggregation (or bagging) — to recover drift-free sensor readings from a suite of …

Evaluation Of Generative Models For Predicting Microstructure Geometries In Laser Powder Bed Fusion Additive Manufacturing, Andy Ramlatchan

Computer Science Theses & Dissertations

In-situ process monitoring for metals additive manufacturing is paramount to the successful build of an object for application in extreme or high stress environments. In selective laser melting additive manufacturing, the process by which a laser melts metal powder during the build will dictate the internal microstructure of that object once the metal cools and solidifies. The difficulty lies in that obtaining enough variety of data to quantify the internal microstructures for the evaluation of its physical properties is problematic, as the laser passes at high speeds over powder grains at a micrometer scale. Imaging the process in-situ is complex …

Adaptive Risk Network Dependency Analysis Of Complex Hierarchical Systems, Katherine L. Smith

Computational Modeling & Simulation Engineering Theses & Dissertations

Recently the number, variety, and complexity of interconnected systems have been increasing while the resources available to increase resilience of those systems have been decreasing. Therefore, it has become increasingly important to quantify the effects of risks and the resulting disruptions over time as they ripple through networks of systems. This dissertation presents a novel modeling and simulation methodology which quantifies resilience, as impact on performance over time, and risk, as the impact of probabilistic disruptions. This work includes four major contributions over the state-of-the-art which are: (1) cyclic dependencies are captured by separation of performance variables into layers which …

Applied Deep Learning: Case Studies In Computer Vision And Natural Language Processing, Md Reshad Ul Hoque

Electrical & Computer Engineering Theses & Dissertations

Deep learning has proved to be successful for many computer vision and natural language processing applications. In this dissertation, three studies have been conducted to show the efficacy of deep learning models for computer vision and natural language processing. In the first study, an efficient deep learning model was proposed for seagrass scar detection in multispectral images which produced robust, accurate scars mappings. In the second study, an arithmetic deep learning model was developed to fuse multi-spectral images collected at different times with different resolutions to generate high-resolution images for downstream tasks including change detection, object detection, and land cover …

Emotion Detection Using An Ensemble Model Trained With Physiological Signals And Inferred Arousal-Valence States, Matthew Nathanael Gray

Electrical & Computer Engineering Theses & Dissertations

Affective computing is an exciting and transformative field that is gaining in popularity among psychologists, statisticians, and computer scientists. The ability of a machine to infer human emotion and mood, i.e. affective states, has the potential to greatly improve human-machine interaction in our increasingly digital world. In this work, an ensemble model methodology for detecting human emotions across multiple subjects is outlined. The Continuously Annotated Signals of Emotion (CASE) dataset, which is a dataset of physiological signals labeled with discrete emotions from video stimuli as well as subject-reported continuous emotions, arousal and valence, from the circumplex model, is used for …

Artificial Intelligence In Financial Technology, Keng Siau, Fiona Fui-Hoon Nah, Brenda L. Eschenbrenner, Langtao Chen

Research Collection School Of Computing and Information Systems

AI applications in health care, communications, and arts have brought about rapid and dramatic advances in these fields. Nevertheless, the rapidly expanding potential of AI in the economy and society has raised a set of challenging issues. The fields of AI and financial technology are not spared. How can artificial intelligence (AI) be utilized in financial technology (fintech)? What will be the impact? What actionable objectives are needed to realize value from AI? This research uses a systematic qualitative research methodology, Value-Focused Thinking, to identify the actionable objectives for deriving value from AI in the fintech industry. The results of …

Crosscbr: Cross‑View Contrastive Learning For Bundle Recommendation, Yunshan Ma, Yingzhi He, An Zhang, Xiang Wang, Tat-Seng Chua

Research Collection School Of Computing and Information Systems

Bundle recommendation aims to recommend a bundle of related items to users, which can satisfy the users' various needs with one-stop convenience. Recent methods usually take advantage of both user-bundle and user-item interactions information to obtain informative representations for users and bundles, corresponding to bundle view and item view, respectively. However, they either use a unified view without differentiation or loosely combine the predictions of two separate views, while the crucial cooperative association between the two views' representations is overlooked.In this work, we propose to model the cooperative association between the two different views through cross-view contrastive learning. By encouraging …

Is Artificial Intelligence A Double-Edged Sword? Insights From Three Essays On Its Impacts, Ankur Arora

Graduate Theses and Dissertations

"If we do it right, we might be able to evolve a form of work that taps into our uniquely human capabilities and restores our humanity. The ultimate paradox is that this technology may become a powerful catalyst that we need to reclaim our humanity." - John Hagel

Artificial intelligence (AI) is viewed as a disruptive technology that some executives believe will take over a lot of jobs. However, others believe that AI will bolster growth, improve business processes, and create new business opportunities. This dissertation focuses on the tension arising from such contrasting expected impacts of AI. Extant research …

Scheduling, Complexity, And Solution Methods For Space Robot On-Orbit Servicing, Susan E. Sorenson

Graduate Theses and Dissertations

This research proposes problems, models, and solutions for the scheduling of space robot on-orbit servicing. We present the Multi-Orbit Routing and Scheduling of Refuellable On-Orbit Servicing Space Robots problem which considers on-orbit servicing across multiple orbits with moving tasks and moving refuelling depots. We formulate a mixed integer linear program model to optimize the routing and scheduling of robot servicers to accomplish on-orbit servicing tasks. We develop and demonstrate flexible algorithms for the creation of the model parameters and associated data sets. Our first algorithm creates the network arcs using orbital mechanics. We have also created a novel way to …

Deep Learning Applications In Industrial And Systems Engineering, Winthrop Harvey

Graduate Theses and Dissertations

Deep learning - the use of large neural networks to perform machine learning - has transformed the world. As the capabilities of deep models continue to grow, deep learning is becoming an increasingly valuable and practical tool for industrial engineering. With its wide applicability, deep learning can be turned to many industrial engineering tasks, including optimization, heuristic search, and functional approximation. In this dissertation, the major concepts and paradigms of deep learning are reviewed, and three industrial engineering projects applying these methods are described. The first applies a deep convolutional network to the task of absolute aerial geolocalization - the …

Meta-Detr: Image-Level Few-Shot Detection With Inter-Class Correlation Exploitation, Gongjie Zhang, Zhipeng Luo, Kaiwen Cui, Shijian Lu, Eric P. Xing

Machine Learning Faculty Publications

Few-shot object detection has been extensively investigated by incorporating meta-learning into region-based detection frameworks. Despite its success, the said paradigm is still constrained by several factors, such as (i) low-quality region proposals for novel classes and (ii) negligence of the inter-class correlation among different classes. Such limitations hinder the generalization of base-class knowledge for the detection of novel-class objects. In this work, we design Meta-DETR, which (i) is the first image-level few-shot detector, and (ii) introduces a novel inter-class correlational meta-learning strategy to capture and leverage the correlation among different classes for robust and accurate few-shot object detection. Meta-DETR works …

Semantic-Aligned Matching For Enhanced Detr Convergence And Multi-Scale Feature Fusion, Gongjie Zhang, Zhipeng Luo, Yingchen Yu, Jiaxing Huang, Kaiwen Cui, Shijian Lu, Eric Xing

Machine Learning Faculty Publications

The recently proposed DEtection TRansformer (DETR) has established a fully end-to-end paradigm for object detection. However, DETR suffers from slow training convergence, which hinders its applicability to various detection tasks. We observe that DETR's slow convergence is largely attributed to the difficulty in matching object queries to relevant regions due to the unaligned semantics between object queries and encoded image features. With this observation, we design Semantic-Aligned-Matching DETR++ (SAM-DETR++) to accelerate DETR's convergence and improve detection performance. The core of SAM-DETR++ is a plug-andplay module that projects object queries and encoded image features into the same feature embedding space, where …

Application In Medicine: Has Artificial Intelligence Stood The Test Of Time, Mir Ibrahim Sajid, Shaheer Ahmed, Usama Waqar, Javeria Tariq, Mohsin Chundrigar, Samira Shabbir Balouch, Sajid Abaidullah

Medical College Documents

Artificial intelligence (AI) has proven time and time again to be a game-changer innovation in every walk of life, including medicine. Introduced by Dr. Gunn in 1976 to accurately diagnose acute abdominal pain and list potential differentials, AI has since come a long way. In particular, AI has been aiding in radiological diagnoses with good sensitivity and specificity by using machine learning algorithms. With the coronavirus disease 2019 pandemic, AI has proven to be more than just a tool to facilitate healthcare workers in decision making and limiting physician-patient contact during the pandemic. It has guided governments and key policymakers …

Towards Smart City Security: Violence And Weaponized Violence Detection Using Dcnn, Toluwani Aremu, Li Zhiyuan, Reem Alameeri, Moayad Aloqaily, Mohsen Guizani

Machine Learning Faculty Publications

In this ever connected society, CCTVs have had a pivotal role in enforcing safety and security of the citizens by recording unlawful activities for the authorities to take actions. In a smart city context, using Deep Convolutional Neural Networks (DCNN) to detection violence and weaponized violence from CCTV videos will provide an additional layer of security by ensuring real-time detection around the clock. In this work, we introduced a new specialised dataset by gathering real CCTV footage of both weaponized and non-weaponized violence as well as non-violence videos from YouTube. We also proposed a novel approach in merging consecutive video …

Computational Models To Detect Radiation In Urban Environments: An Application Of Signal Processing Techniques And Neural Networks To Radiation Data Analysis, Jose Nicolas Gachancipa

Beyond: Undergraduate Research Journal

Radioactive sources, such as uranium-235, are nuclides that emit ionizing radiation, and which can be used to build nuclear weapons. In public areas, the presence of a radioactive nuclide can present a risk to the population, and therefore, it is imperative that threats are identified by radiological search and response teams in a timely and effective manner. In urban environments, such as densely populated cities, radioactive sources may be more difficult to detect, since background radiation produced by surrounding objects and structures (e.g., buildings, cars) can hinder the effective detection of unnatural radioactive material. This article presents a computational model …

Self-Distilled Vision Transformer For Domain Generalization, Maryam Sultana, Muzammal Naseer, Muhammad Haris Khan, Salman Khan, Fahad Shahbaz Khan

Computer Vision Faculty Publications

In recent past, several domain generalization (DG) methods have been proposed, showing encouraging performance, however, almost all of them build on convolutional neural networks (CNNs). There is little to no progress on studying the DG performance of vision transformers (ViTs), which are challenging the supremacy of CNNs on standard benchmarks, often built on i.i.d assumption. This renders the real-world deployment of ViTs doubtful. In this paper, we attempt to explore ViTs towards addressing the DG problem. Similar to CNNs, ViTs also struggle in out-of-distribution scenarios and the main culprit is overfitting to source domains. Inspired by the modular architecture of …

Reinforcement Actor-Critic Learning As A Rehearsal In Microrts, Shiron Manandhar

Master's Theses

Real-time strategy (RTS) games have provided a fertile ground for AI research with notable recent successes based on deep reinforcement learning (RL). However, RL remains a data-hungry approach featuring a high sample complexity. In this thesis, we focus on a sample complexity reduction technique called reinforcement learning as a rehearsal (RLaR), and on the RTS game of MicroRTS to formulate and evaluate it. RLaR has been formulated in the context of action-value function based RL before. Here we formulate it for a different RL framework, called actor-critic RL. We show that on the one hand the actor-critic framework allows RLaR …

A Modified Point Matching Method For Non-Rigid Image Registration, Jintai Shangguan, Yawen Dang, Wei Lian

Journal of System Simulation

Abstract: Aiming at the problem that the registration results tend to converge to local minima due to the complexity of the relative position changes between two point sets in the non-rigid body point matching process, a joint estimation method for non-rigid body point matching based on precenter alignment is proposed, a modified matching method for non-rigid image registration based on centre preregistration is proposed. To better achieve the point matching accuracy between two point sets, a centre preregistration step is applied before the iterative closest point matching algorithm, which converges to a solution more close to a global optimum and …

Transfer Method Of Operational Simulation Experiment Scope Using Compromised Case-Based Reasoning, Jun Ma, Jingyu Yang, Xi Wu

Journal of System Simulation

Abstract: The scope of operational simulation experiment is usually determined by experts, which costs relatively high. In order to transfer the knowledge of experimental scope selection from historical data of operational simulation experiment to new operational experiment cases, the method of compromised case-based reasoning is proposed. According to the data characteristics of the case, the representation method of the operational simulation experiment case is proposed; according to the structure and attribute characteristics of the case, the hybrid similarity calculation method of subjective and objective comprehensive weighting is proposed; aiming at the problems of retrieval failure and less information content …

Optimal Dispatch Of Integrated Energy System Considering Ladder-Type Carbon Trading, Liying Wang, Jialin Lin, Houqi Dong, Ming Zeng, Yuqing Wang

Journal of System Simulation

Abstract: With the development of the electricity market and carbon market,the introduction of demand response and carbon trading mechanisms into the operation and dispatch of integrated energy systems will help guide users and system operators to optimize electricity consumption and dispatch plans.The comprehensive incentive measures such as time-of-use electricity prices and demand response incentive subsidies are used to guide users to participate in demand response.A two-layer stochastic optimal scheduling model for a comprehensive energy system considering the ladder-type carbon trading mechanism and demand response is constructed based on IGDT (information gap decision theory) theory.The two-layer model is converted …

Bi-Level Optimization Of Distribution Network For Hybrid Energy Storage System Of Storage Battery And Hydrogen Storage, Feibo Feng, Xingde Yan, Baoqiang Zheng, Xiaofeng Yin, Mengzhen Zhou, Xin Jiang

Journal of System Simulation

Abstract: Under the background of carbon neutralization and emission peaking goals and the utilization of clean hydrogen energy, aiming at the demand of distribution network configuring electrochemical energy storage and hydrogen energy storage system to form a hybrid energy storage system to improve power quality, a bi-level optimization model of the hybrid energy storage system is established. The upper level location and capacity model comprehensively considers the investment cost, network loss cost and voltage offset, while the lower level optimization operation model considers the operation cost of hybrid energy storage system, and the voltage stability index is introduced for evaluation. …

Modeling And Simulation Of Optimal Strategy For Electric Vehicles Participating In Power Grid Frequency Regulation, Li Yao, Junjie Hu, Wenshuai Ma, Zhile Yang

Journal of System Simulation

Abstract: Electric vehicles (EVs) have similar characteristics of distributed energy storage, and making full use of the flexibility of EVs can provide ancillary services to the grid and gain benefits. Considering the influence of uncertain factors, a bidding model for electric vehicle aggregator (EVA) to participate in the day-ahead energy market and frequency regulation ancillary service market is constructed with the maximum revenue expectation of EVA as the target. A real-time energy distribution incentive strategy based on contract theory is proposed to realize the distribution of EVA's frequency regulation demand under the condition of maximizing social welfare. Through case studies, …

Simulation Model Of Forest Fire Spread Based On Swarm Intelligence, Aibin Chen, Fubo Ding, Guoxiong Zhou, Bo Zhou

Journal of System Simulation

Abstract: Aiming at the shortcomings of high computational complexity and low simulation accuracy of traditional forest fire spread model, a forest fire spread simulation model based on swarm intelligence is proposed.By establishing fuel factor matrix and landform factor matrix, and combining with the real-time meteorological information, the computational complexity is reduced; the spread behavior of the forest fire is abstracted as the cluster behavior of each module fire point, and the correlation between modules is considered to improve the accuracy of forest fire spread simulation model.The model is compared with Wang Zhengfei model and two-dimensional cellular automata model. …

Two Stage Optimization Algorithm To Solve The Green Packing Vehicle Routing Problem, Rong Hu, Wen Jiang, Bin Qian, Naikang Yu

Journal of System Simulation

Abstract: The green open vehicle routing problem with two-dimensional loading constraints (2L-GOVRP) is integration of the green open vehicle routing problem and two-dimensional bin packing problem. The model of 2L-GOVRP is established and a two-stage optimization algorithm (TSOA) is proposed to minimize fuel consumption. In the first stage of TSOA, adaptive whale optimization algorithm (AWOA) is designed to solve the vehicle routing problem, which determine the initial delivery route of the vehicle (the initial solution of 2L-GOVRP). The algorithm has four kinds of variable neighborhoods local operation to perform a local search. In the second stage of TSOA, the skyline …

A Method Of Loose Coupling Entity Modeling Based On Variable Rules, Zheng Yang, Zhimin Xiang, Shiwen Ma

Journal of System Simulation

Abstract: Operational Entity Modeling is a hot research topic in the field of combat simulation. A loose coupling entity modeling method based on variable rules is proposed. The architecture of operational entity model based on variable rules and the internal and external interaction mechanism of the model are presented in terms of entity, mission, action, interaction, event and rule. On this basis, the running framework of operational entity model is designed, and the entity model uniform scheduling mechanism is standardized, which solves the problems of over-tight coupling of operational rules in the operational entity model and low reliability of the …

Research On Prediction Of Model Based On Multi-Scale Lstm, Junjie Qiu, Hong Zheng, Yunhui Cheng

Journal of System Simulation

Abstract: Aircraft engine remaining useful life (RUL) prediction is the core issue in equipmentfailure prognostics and health management (PHM). Aiming at the characteristics of high dimensionality, high lag and complexity of engine data, a multi-scale attention-based bidirectional long short-term memory neural network model based on self-training weights is proposed. Multi-scale features are extracted through bidirectional long short-term memory neural network (BiLSTM) of different scales. A fusion algorithm based on self-training weights is proposed, and an attention mechanism is introduced to screen features at different scales to improve prediction accuracy. Various models are compared on the NASA's C-MAPSS data set. The …