Artificial Intelligence and Robotics Commons^™

Reinforcement Learning With Deep Q-Networks, Caleb Cassady

Masters Theses & Specialist Projects

In the past decade, machine learning strategies centered on the use of Deep Neural Networks (DNNs) have caught the interest of researchers due to their success in complicated classification and prediction problems. More recently, these DNNs have been applied to reinforcement learning tasks with state of- the-art results using Deep Q-Networks (DQNs) based on the Q-Learning algorithm. However, the DQN training process is different from standard DNNs and poses significant challenges for certain reinforcement learning environments. This paper examines some of these challenges, compares proposed solutions, and offers novel solutions based on previous research. Experiment implementation available at https://github.com/caleb98/dqlearning.

Comai: Enabling Lightweight, Collaborative Intelligence By Retrofitting Vision Dnns, Kasthuri Jayarajah, Dhanuja Wanniarachchige, Tarek Abdelzaher, Archan Misra

Research Collection School Of Computing and Information Systems

While Deep Neural Network (DNN) models have transformed machine vision capabilities, their extremely high computational complexity and model sizes present a formidable deployment roadblock for AIoT applications. We show that the complexity-vs-accuracy-vs-communication tradeoffs for such DNN models can be significantly addressed via a novel, lightweight form of “collaborative machine intelligence” that requires only runtime changes to the inference process. In our proposed approach, called ComAI, the DNN pipelines of different vision sensors share intermediate processing state with one another, effectively providing hints about objects located within their mutually-overlapping Field-of-Views (FoVs). CoMAI uses two novel techniques: (a) a secondary shallow ML …

Understanding Deep Learning - Challenges And Prospects, Niha Adnan, Fahad Umer

Department of Surgery Docs

The developments in Artificial Intelligence have been on the rise since its advent. The advancements in this field have been the innovative research area across a wide range of industries, making its incorporation in dentistry inevitable. Artificial Intelligence techniques are making serious progress in the diagnostic and treatment planning aspects of dental clinical practice. This will ultimately help in the elimination of subjectivity and human error that are often part of radiographic interpretations, and will improve the overall efficiency of the process. The various types of Artificial Intelligence algorithms that exist today make the understanding of their application quite complex. …

Deeply Learning Deep Inelastic Scattering Kinematics, Markus Diefenthaler, Abdullah Farhat, Andrii Verbytskyi, Yuesheng Xu

Mathematics & Statistics Faculty Publications

We study the use of deep learning techniques to reconstruct the kinematics of the neutral current deep inelastic scattering (DIS) process in electron–proton collisions. In particular, we use simulated data from the ZEUS experiment at the HERA accelerator facility, and train deep neural networks to reconstruct the kinematic variables Q² and x. Our approach is based on the information used in the classical construction methods, the measurements of the scattered lepton, and the hadronic final state in the detector, but is enhanced through correlations and patterns revealed with the simulated data sets. We show that, with the appropriate selection …

Automatic Cephalometric Landmark Detection On X-Ray Images Using Object Detection, Cheng-Ho King, Yin-Lin Wang, Chia-Ling Tsai

Publications and Research

We propose a new deep convolutional cephalometric landmark detection framework for orthodontic treatment. Our proposed method consists of two major steps: landmark detection using a deep neural network for object detection, and landmark repair to ensure one instance per landmark class. For landmark detection, we modify the loss function of the backbone network YOLOv3 to eliminate the constrains on the bounding box and incorporate attention mechanism to improve the detection accuracy. For landmark repair, a triangle mesh is generated from the average face to eliminate superfluous instances, followed by estimation of missing landmarks from the detected ones using Laplacian Mesh. …

Security Hardening Of Intelligent Reflecting Surfaces Against Adversarial Machine Learning Attacks, Ferhat Ozgur Catak, Murat Kuzlu, Haolin Tang, Evren Catak, Yanxiao Zhao

Engineering Technology Faculty Publications

Next-generation communication networks, also known as NextG or 5G and beyond, are the future data transmission systems that aim to connect a large amount of Internet of Things (IoT) devices, systems, applications, and consumers at high-speed data transmission and low latency. Fortunately, NextG networks can achieve these goals with advanced telecommunication, computing, and Artificial Intelligence (AI) technologies in the last decades and support a wide range of new applications. Among advanced technologies, AI has a significant and unique contribution to achieving these goals for beamforming, channel estimation, and Intelligent Reflecting Surfaces (IRS) applications of 5G and beyond networks. However, the …

Arithfusion: An Arithmetic Deep Model For Temporal Remote Sensing Image Fusion, Md Reshad Ul Hoque, Jian Wu, Chiman Kwan, Krzysztof Koperski, Jiang Li

Electrical & Computer Engineering Faculty Publications

Different satellite images may consist of variable numbers of channels which have different resolutions, and each satellite has a unique revisit period. For example, the Landsat-8 satellite images have 30 m resolution in their multispectral channels, the Sentinel-2 satellite images have 10 m resolution in the pan-sharp channel, and the National Agriculture Imagery Program (NAIP) aerial images have 1 m resolution. In this study, we propose a simple yet effective arithmetic deep model for multimodal temporal remote sensing image fusion. The proposed model takes both low- and high-resolution remote sensing images at t₁ together with low-resolution images at a …

Machine Learning And Computer Vision In Solar Physics, Haodi Jiang

Dissertations

In the recent decades, the difficult task of understanding and predicting violent solar eruptions and their terrestrial impacts has become a strategic national priority, as it affects the life of human beings, including communication, transportation, the power grid, national defense, space travel, and more. This dissertation explores new machine learning and computer vision techniques to tackle this difficult task. Specifically, the dissertation addresses four interrelated problems in solar physics: magnetic flux tracking, fibril tracing, Stokes inversion and vector magnetogram generation.

First, the dissertation presents a new deep learning method, named SolarUnet, to identify and track solar magnetic flux elements in …

Characterizing Convolutional Neural Network Early-Learning And Accelerating Non-Adaptive, First-Order Methods With Localized Lagrangian Restricted Memory Level Bundling, Benjamin O. Morris

Theses and Dissertations

This dissertation studies the underlying optimization problem encountered during the early-learning stages of convolutional neural networks and introduces a training algorithm competitive with existing state-of-the-art methods. First, a Design of Experiments method is introduced to systematically measure empirical second-order Lipschitz upper bound and region size estimates for local regions of convolutional neural network loss surfaces experienced during the early-learning stages. This method demonstrates that architecture choices can significantly impact the local loss surfaces traversed during training. Next, a Design of Experiments method is used to study the effects convolutional neural network architecture hyperparameters have on different optimization routines' abilities to …

Splash: Learnable Activation Functions For Improving Accuracy And Adversarial Robustness, Mohammadamin Tavakoli, Forest Agostinelli, Pierre Baldi

Publications

We introduce SPLASH units, a class of learnable activation functions shown to simultaneously improve the accuracy of deep neural networks while also improving their robustness to adversarial attacks. SPLASH units have both a simple parameterization and maintain the ability to approximate a wide range of non-linear functions. SPLASH units are: (1) continuous; (2) grounded (f(0)=0"); (3) use symmetric hinges; and (4) their hinges are placed at fixed locations which are derived from the data (i.e. no learning required). Compared to nine other learned and fixed activation functions, including ReLU and its variants, SPLASH units show superior performance across three datasets …

Flying Free: A Research Overview Of Deep Learning In Drone Navigation Autonomy, Thomas Lee, Susan Mckeever, Jane Courtney

Articles

With the rise of Deep Learning approaches in computer vision applications, significant strides have been made towards vehicular autonomy. Research activity in autonomous drone navigation has increased rapidly in the past five years, and drones are moving fast towards the ultimate goal of near-complete autonomy. However, while much work in the area focuses on specific tasks in drone navigation, the contribution to the overall goal of autonomy is often not assessed, and a comprehensive overview is needed. In this work, a taxonomy of drone navigation autonomy is established by mapping the definitions of vehicular autonomy levels, as defined by the …

Year-Independent Prediction Of Food Insecurity Using Classical & Neural Network Machine Learning Methods, Caleb Christiansen, Torrey J. Wagner, Brent Langhals

Faculty Publications

Current food crisis predictions are developed by the Famine Early Warning System Network, but they fail to classify the majority of food crisis outbreaks with model metrics of recall (0.23), precision (0.42), and f1 (0.30). In this work, using a World Bank dataset, classical and neural network (NN) machine learning algorithms were developed to predict food crises in 21 countries. The best classical logistic regression algorithm achieved a high level of significance (p < 0.001) and precision (0.75) but was deficient in recall (0.20) and f1 (0.32). Of particular interest, the classical algorithm indicated that the vegetation index and the food price index were both positively correlated with food crises. A novel method for performing an iterative multidimensional hyperparameter search is presented, which resulted in significantly improved performance when applied to this dataset. Four iterations were conducted, which resulted in excellent 0.96 for metrics of precision, recall, and f1. Due to this strong performance, the food crisis year was removed from the dataset to prevent immediate extrapolation when used on future data, and the modeling process was repeated. The best “no year” model metrics remained strong, achieving ≥0.92 for recall, precision, and f1 while meeting a 10% f1 overfitting threshold on the test (0.84) and holdout (0.83) datasets. The year-agnostic neural network model represents a novel approach to classify food crises and outperforms current food crisis prediction efforts.

Using Deep Learning To Analyze Materials In Medical Images, Carson Molder

Computer Science and Computer Engineering Undergraduate Honors Theses

Modern deep learning architectures have become increasingly popular in medicine, especially for analyzing medical images. In some medical applications, deep learning image analysis models have been more accurate at predicting medical conditions than experts. Deep learning has also been effective for material analysis on photographs. We aim to leverage deep learning to perform material analysis on medical images. Because material datasets for medicine are scarce, we first introduce a texture dataset generation algorithm that automatically samples desired textures from annotated or unannotated medical images. Second, we use a novel Siamese neural network called D-CNN to predict patch similarity and build …

Convolutional Audio Source Separation Applied To Drum Signal Separation, Marius Orehovschi

Honors Theses

This study examined the task of drum signal separation from full music mixes via both classical methods (Independent Component Analysis) and a combination of Time-Frequency Binary Masking and Convolutional Neural Networks. The results indicate that classical methods relying on predefined computations do not achieve any meaningful results, while convolutional neural networks can achieve imperfect but musically useful results. Furthermore, neural network performance can be improved by data augmentation via transposition – a technique that can only be applied in the context of drum signal separation.

Reliable And Interpretable Machine Learning For Modeling Physical And Cyber Systems, Daniel L. Marino Lizarazo

Theses and Dissertations

Over the past decade, Machine Learning (ML) research has predominantly focused on building extremely complex models in order to improve predictive performance. The idea was that performance can be improved by adding complexity to the models. This approach proved to be successful in creating models that can approximate highly complex relationships while taking advantage of large datasets. However, this approach led to extremely complex black-box models that lack reliability and are difficult to interpret. By lack of reliability, we specifically refer to the lack of consistent (unpredictable) behavior in situations outside the training data. Lack of interpretability refers to the …

Stock Trend Prediction Using Candlestick Charting And Ensemble Machine Learning Techniques With A Novelty Feature Engineering Scheme, Yaohu Lin, Shancun Liu, Haijun Yang, Harris Wu

Information Technology & Decision Sciences Faculty Publications

Stock market forecasting is a knotty challenging task due to the highly noisy, nonparametric, complex and chaotic nature of the stock price time series. With a simple eight-trigram feature engineering scheme of the inter-day candlestick patterns, we construct a novel ensemble machine learning framework for daily stock pattern prediction, combining traditional candlestick charting with the latest artificial intelligence methods. Several machine learning techniques, including deep learning methods, are applied to stock data to predict the direction of the closing price. This framework can give a suitable machine learning prediction method for each pattern based on the trained results. The investment …

A Tool For Segmentation Of Secondary Structures In 3d Cryo-Em Density Map Components Using Deep Convolutional Neural Networks, Yongcheng Mu, Salim Sazzed, Maytha Alshammari, Jiangwen Sun, Jing He

Computer Science Faculty Publications

Although cryo-electron microscopy (cryo-EM) has been successfully used to derive atomic structures for many proteins, it is still challenging to derive atomic structures when the resolution of cryo-EM density maps is in the medium resolution range, such as 5–10 Å. Detection of protein secondary structures, such as helices and β-sheets, from cryo-EM density maps provides constraints for deriving atomic structures from such maps. As more deep learning methodologies are being developed for solving various molecular problems, effective tools are needed for users to access them. We have developed an effective software bundle, DeepSSETracer, for the detection of protein secondary structure …

Large Scale Subject Category Classification Of Scholarly Papers With Deep Attentive Neural Networks, Bharath Kandimalla, Shaurya Rohatgi, Jian Wu, C. Lee Giles

Computer Science Faculty Publications

Subject categories of scholarly papers generally refer to the knowledge domain(s) to which the papers belong, examples being computer science or physics. Subject category classification is a prerequisite for bibliometric studies, organizing scientific publications for domain knowledge extraction, and facilitating faceted searches for digital library search engines. Unfortunately, many academic papers do not have such information as part of their metadata. Most existing methods for solving this task focus on unsupervised learning that often relies on citation networks. However, a complete list of papers citing the current paper may not be readily available. In particular, new papers that have few …

Computational Cognition And Deep Learning, Andy Malinsky

The Compass

No abstract provided.

Modular Neural Networks For Low-Power Image Classification On Embedded Devices, Abhinav Goel, Sara Aghajanzadeh, Caleb Tung, Shuo-Han Chen, George K. Thiruvathukal, Yung-Hisang Lu

Computer Science: Faculty Publications and Other Works

Embedded devices are generally small, battery-powered computers with limited hardware resources. It is difficult to run deep neural networks (DNNs) on these devices, because DNNs perform millions of operations and consume significant amounts of energy. Prior research has shown that a considerable number of a DNN’s memory accesses and computation are redundant when performing tasks like image classification. To reduce this redundancy and thereby reduce the energy consumption of DNNs, we introduce the Modular Neural Network Tree architecture. Instead of using one large DNN for the classifier, this architecture uses multiple smaller DNNs (called modules) to progressively classify images …

Fault Diagnosis For Automobile Coating Equipments Based On Extension Neural Network, Yongwei Ye, Shedong Ren, Lianqiang Ye, Shenhao Ge, Zhiqin Qian

Journal of System Simulation

Abstract: Aiming at the difficulty in discovering and eliminating the system faults of automobile coating equipments in time, a new method of fault diagnosis based on extension neural network was proposed. The feature of extension theory was used in managing the structured information through qualitative and quantitative description, and it was also combined by the characteristic of parallel construct in neural network. So the extension reasoning process was completed by means of the parallel distributed processing construct of the network. Matter-element input and output models were established according to the equipment monitoring parameters and fault types for the heating system. …

Research On Asymptotic Stability For Markovian Jumping Neural Network With Unknown Transition Probabilities, Lu Yang, Shujuan Yi, Weijian Ren, Jiandong Liu

Journal of System Simulation

Abstract: The analysis problem of asymptotic stability for a class of uncertain neural networks with Markovian jumping parameters and time delays was addressed. The general representative dynamic stochastic neural network model was established. The considered transition probabilities were assumed to be partially unknown. The parameter uncertainties were considered to be norm-bounded. Based on Lyapunov stability theory, by constructing a suitable Lyapunov-Krasovskii function and using the stochastic analysis method, some sufficient criteria for the stability of discrete Markovian neural networks was derived. Through the Matlab LMI toolbox, solving a set of linear matrix inequalities to test criterion, the new criterion reduced …

Using Generative Adversarial Networks To Classify Structural Damage Caused By Earthquakes, Gian P. Delacruz

Master's Theses

The amount of structural damage image data produced in the aftermath of an earthquake can be staggering. It is challenging for a few human volunteers to efficiently filter and tag these images with meaningful damage information. There are several solution to automate post-earthquake reconnaissance image tagging using Machine Learning (ML) solutions to classify each occurrence of damage per building material and structural member type. ML algorithms are data driven; improving with increased training data. Thanks to the vast amount of data available and advances in computer architectures, ML and in particular Deep Learning (DL) has become one of the most …

Research On Image Description Method Based On Neural Network, Kong Rui, Xie Wei, Lei Tai

Journal of System Simulation

Abstract: The automatic recognition and automatically describing image content is an important research direction to the artificial intelligence to connect the computer vision and the natural language processing. A method of describing the image content is proposed to generate the natural language by using the deep neural network model. The model consists of a convolutional neural network (CNN) and a recurrent neural network (RNN). The CNN is used to extract features of the input image to generate a fixed-length feature vector, which initializes the RNN to generate the sentences. Experimental results on the MSCOCO image description dataset show the syntactic …

A Collaboration Between Neural Networks And Reinforcement Learning: Applying Concepts To A Brick Breaking Game, Bryce Kadrlik

Augsburg Honors Review

The intent of this work is to explore the interactions of artificial neural networks and digital games. It details the development of an artificial neural network trained upon a brick breaking game like the Atari game Breakout. This network was designed with the goals of not dropping the ball and maximizing the game score. Full game and network integration was not completed. However, two versions of the network were developed to move the paddle to the right or left based on the ball's point of impact on the paddle. In preliminary testing using manual inputs, these networks eventually learned to …

Final Presentation To The Library Of Congress On Digital Libraries, Intelligent Data Analytics, And Augmented Description, Elizabeth Lorang, Leen-Kiat Soh, Yi Liu, Chulwoo Pack

University of Nebraska-Lincoln Libraries: Presentations

This presentation to Library of Congress staff, delivered onsite on January 10, 2020, presents a tour through the demonstration project pursued by the Aida digital libraries research team with the Library of Congress in 2019-2020. In addition to providing an overview and analysis of the specific machine learning projects scoped and explored, this presentation includes a number of high-level take-aways and recommendations designed to influence and inform the Library of Congress's machine learning efforts going forward.

Digital Libraries, Intelligent Data Analytics, And Augmented Description: A Demonstration Project, Elizabeth Lorang, Leen-Kiat Soh, Yi Liu, Chulwoo Pack

University of Nebraska-Lincoln Libraries: Faculty Publications

From July 16-to November 8, 2019, the Aida digital libraries research team at the University of Nebraska-Lincoln collaborated with the Library of Congress on “Digital Libraries, Intelligent Data Analytics, and Augmented Description: A Demonstration Project.“ This demonstration project sought to (1) develop and investigate the viability and feasibility of textual and image-based data analytics approaches to support and facilitate discovery; (2) understand technical tools and requirements for the Library of Congress to improve access and discovery of its digital collections; and (3) enable the Library of Congress to plan for future possibilities. In pursuit of these goals, we focused our …

A Machine Learning System For Glaucoma Detection Using Inexpensive Machine Learning, Jon Kilgannon

West Chester University Master’s Theses

This thesis presents a neural network system which segments images of the retina to calculate the cup-to-disc ratio, one of the diagnostic indicators of the presence or continuing development of glaucoma, a disease of the eye which causes blindness. The neural network is designed to run on commodity hardware and to be run with minimal skill required from the user by packaging the software required to run the network into a Singularity image. The RIGA dataset used to train the network provides images of the retina which have been annotated with the location of the optic cup and disc by …

A Deep Learning Approach To Mapping Irrigation: U-Net Irrmapper, Thomas Henry Colligan Iv

Graduate Student Theses, Dissertations, & Professional Papers

Accurate maps of irrigation are essential for understanding and managing water resources in light of a warming climate. We present a new method for mapping irrigation and apply it to the state of Montana over the years 2000-2019. The method is based on an ensemble of convolutional neural networks that only rely on raw Landsat surface reflectance data. The ensemble of networks method learns to mask clouds and ignore Landsat 7 scan-line failures without supervision, reducing the need for preprocessing data or feature engineering. Unlike other approaches to mapping irrigation, the method doesn't use other mapping products like the Cropland …

Image Features For Tuberculosis Classification In Digital Chest Radiographs, Brian Hooper

All Master's Theses

Tuberculosis (TB) is a respiratory disease which affects millions of people each year, accounting for the tenth leading cause of death worldwide, and is especially prevalent in underdeveloped regions where access to adequate medical care may be limited. Analysis of digital chest radiographs (CXRs) is a common and inexpensive method for the diagnosis of TB; however, a trained radiologist is required to interpret the results, and is subject to human error. Computer-Aided Detection (CAD) systems are a promising machine-learning based solution to automate the diagnosis of TB from CXR images. As the dimensionality of a high-resolution CXR image is very …