Artificial Intelligence and Robotics Commons^™

Computer-Aided Diagnosis Of Low Grade Endometrial Stromal Sarcoma (Lgess), Xinxin Yang, Mark Stamp

Faculty Research, Scholarly, and Creative Activity

Low grade endometrial stromal sarcoma (LGESS) accounts for about 0.2% of all uterine cancer cases. Approximately 75% of LGESS patients are initially misdiagnosed with leiomyoma, which is a type of benign tumor, also known as fibroids. In this research, uterine tissue biopsy images of potential LGESS patients are preprocessed using segmentation and stain normalization algorithms. We then apply a variety of classic machine learning and advanced deep learning models to classify tissue images as either benign or cancerous. For the classic techniques considered, the highest classification accuracy we attain is about 0.85, while our best deep learning model achieves an …

Advancing Proper Dataset Partitioning And Classification Of Visual Search And The Vigilance Decrement Using Eeg Deep Learning Algorithms, Alexander J. Kamrud

Theses and Dissertations

Electroencephalography (EEG) classification of visual search and vigilance tasks has vast potential in its benefits. In future human-machine teaming systems, EEG could act as the tool for operator state assessment, enabling AI teammates to know when to assist the operator in these tasks, with the potential to lead to increased safety of operations, better training systems for our operators, and improved operational effectiveness. This research investigates deep learning methods which utilize EEG signals to classify the efficiency of an operator's search and to classify whether an operator is in a decrement during a vigilance type task, and investigates performing these …

Gradient Free Sign Activation Zero One Loss Neural Networks For Adversarially Robust Classification, Yunzhe Xue

Dissertations

The zero-one loss function is less sensitive to outliers than convex surrogate losses such as hinge and cross-entropy. However, as a non-convex function, it has a large number of local minima, andits undifferentiable attribute makes it impossible to use backpropagation, a method widely used in training current state-of-the-art neural networks. When zero-one loss is applied to deep neural networks, the entire training process becomes challenging. On the other hand, a massive non-unique solution probably also brings different decision boundaries when optimizing zero-one loss, making it possible to fight against transferable adversarial examples, which is a common weakness in deep learning …

Towards Adversarial Robustness With 01 Lossmodels, And Novel Convolutional Neural Netsystems For Ultrasound Images, Meiyan Xie

Dissertations

This dissertation investigates adversarial robustness with 01 loss models and a novel convolutional neural net systems for vascular ultrasound images.

In the first part, the dissertation presents stochastic coordinate descent for 01 loss and its sensitivity to adversarial attacks. The study here suggests that 01 loss may be more resilient to adversarial attacks than the hinge loss and further work is required.

In the second part, this dissertation proposes sign activation network with a novel gradient-free stochastic coordinate descent algorithm and its ensembling model. The study here finds that the ensembling model gives a high minimum distortion (as measured by …

Advances In Deep Learning With Applications To Computer Vision And Astronomy, Zhihang Hu

Dissertations

Deep Learning has spanned a variety of applications in computer vision as well as computational astronomy. These two aspects obtained similar data structure, therefore, their solutions can be transferable between each other. This dissertation look into two video-related tasks in computer vision and propose a novel problem in computational astronomy.

Specifically, acquiring an in-depth understanding of videos has been a cornerstone problem in computer vision. This problem has been studied by various researchers from different perspectives, among which video prediction has attracted much attention. Video prediction aims to generate the pixels of future frames given a sequence of context frames. …

Identification Of Chemical Structures And Substructures Via Deep Q-Learning And Supervised Learning Of Ftir Spectra, Joshua D. Ellis

Graduate Theses/Dissertations

Fourier-transform infrared (FTIR) spectra of organic compounds can be used to compare and identify compounds. A mid-FTIR spectrum gives absorbance values of a compound over the 400-4000 cm^-1 range. Spectral matching is the process of comparing the spectral signature of two or more compounds and returning a value for the similarity of the compounds based on how closely their spectra match. This process is commonly used to identify an unknown compound by searching for its spectrum’s closes match in a database of known spectra. A major limitation of this process is that it can only be used to identify …

Predicting Severity Of Traumatic Brain Injury: A Residual Learning Model From Magnetic Resonance Images, Dacosta Yeboah

Graduate Theses/Dissertations

One of the most significant frontiers for computational scientists is the engineering of human healthcare delivery based on intelligent analysis of health data. In a variety of neurological disorders such as Traumatic Brain Injury (TBI), neuro-imaging information plays a crucial role in the decision-making regarding patient care and as a potential prognostic marker for outcome. TBI is a heterogeneous neurological disorder. Due to the economic burdens of the disorder, sorting out this heterogeneity could provide more insights and better understanding of TBI recovery trajectories, thus improving overall diagnosis and treatment options. Magnetic Resonance Imaging (MRI) is a non-invasive technique that …

Automation Of Radiation Treatment Planning For Cervical Cancer, Dong Joo Rhee

Dissertations and Theses (Open Access)

Cervical cancer is one of the most common cancer in low- and middle-income countries (LMICs). The mortality rate can be reduced if radiation treatment becomes widely available. However, due to the lack of radiation treatment facilities and human resources, many cervical cancer patients in Africa are not able to receive timely treatments or advanced therapies. To increase the availability of radiation treatment in low-and middle-income countries (LMICs) including African countries, many attempts have been made to reduce the cost of medical linear accelerators. However, increasing the number of treatment machines would not instantly resolve the issues, as there would be …

Applying Deep Learning To The Ice Cream Vendor Problem: An Extension Of The Newsvendor Problem, Gaffar Solihu

Electronic Theses and Dissertations

The Newsvendor problem is a classical supply chain problem used to develop strategies for inventory optimization. The goal of the newsvendor problem is to predict the optimal order quantity of a product to meet an uncertain demand in the future, given that the demand distribution itself is known. The Ice Cream Vendor Problem extends the classical newsvendor problem to an uncertain demand with unknown distribution, albeit a distribution that is known to depend on exogenous features. The goal is thus to estimate the order quantity that minimizes the total cost when demand does not follow any known statistical distribution. The …

A Deep Learning Approach For Forecasting Global Commodities Prices, Ahmed Saied Elberawi, Mohamed Belal Prof.

Future Computing and Informatics Journal

Forecasting future values of time-series data is a critical task in many disciplines including financial planning and decision-making. Researchers and practitioners in statistics apply traditional statistical methods (such as ARMA, ARIMA, ES, and GARCH) for a long time with varying accuracies. Deep learning provides more sophisticated and non-linear approximation that supersede traditional statistical methods in most cases. Deep learning methods require minimal features engineering compared to other methods; it adopts an end-to-end learning methodology. In addition, it can handle a huge amount of data and variables. Financial time series forecasting poses a challenge due to its high volatility and non-stationarity …

Signal Processing And Data Analysis For Real-Time Intermodal Freight Classification Through A Multimodal Sensor System., Enrique J. Sanchez Headley

Graduate Theses and Dissertations

Identifying freight patterns in transit is a common need among commercial and municipal entities. For example, the allocation of resources among Departments of Transportation is often predicated on an understanding of freight patterns along major highways. There exist multiple sensor systems to detect and count vehicles at areas of interest. Many of these sensors are limited in their ability to detect more specific features of vehicles in traffic or are unable to perform well in adverse weather conditions. Despite this limitation, to date there is little comparative analysis among Laser Imaging and Detection and Ranging (LIDAR) sensors for freight detection …

Research On Intrusion Detection Based On Stacked Autoencoder And Long-Short Memory, Lin Shuo, An Lei, Zhijun Gao, Shan Dan, Wenli Shang

Journal of System Simulation

Abstract: As network attacks increasingly hidden, intelligent and complex. Simple machine learning cannot deal with attacks timely. A deep learning method based on the combination of SDAE and LSTM is proposed. Firstly, the distribution rules of network data are extracted intelligently layer by layer by SDAE, and the diverse anomaly features of high-dimensional data ate extracted by using coefficient penalty and reconstruction error of each coding layer. Then, LSTM’ s memory function and the powerful learning ability of sequence data are used to classify learning depth. Finally, the experiments are carried out with the UNSW-NB15 data set, which is analyzed …

Creating Synthetic Satellite Cloud Data Based On Gan Method, Wencong Cheng, Xiaokang Shi, Zhigang Wang

Journal of System Simulation

Abstract: To create the synthetic satellite cloud data in the domain of Meteorology, a method based on Generative Adversarial Networks (GAN) is proposed. Depending on ability of the nonlinear mapping and the information extraction of raster data with the deep learning network, a deep generative adversarial network model is proposed to extract the corresponding information between the numerical weather prediction(NWP) products and the satellite cloud data, and then the appropriate elements of the NWP product are chosen as the input to synthesize the corresponding satellite cloud data. The experiments are conducted on the re-analysis products of the European Centre …

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 …

Grand-Vision: An Intelligent System For Optimized Deployment Scheduling Of Law Enforcement Agents, Jonathan Chase, Tran Phong, Kang Long, Tony Le, Hoong Chuin Lau

Research Collection School Of Computing and Information Systems

Law enforcement agencies in dense urban environments, faced with a wide range of incidents to handle and limited manpower, are turning to data-driven AI to inform their policing strategy. In this paper we present a patrol scheduling system called GRAND-VISION: Ground Response Allocation and Deployment - Visualization, Simulation, and Optimization. The system employs deep learning to generate incident sets that are used to train a patrol schedule that can accommodate varying manpower, break times, manual pre-allocations, and a variety of spatio-temporal demand features. The complexity of the scenario results in a system with real world applicability, which we demonstrate through …

Deep Learning On Image Forensics And Anti-Forensics, Zhangyi Shen

Dissertations

Image forensics protect the authenticity and integrity of digital images. On the contrary, as the countermeasures of digital forensics, anti-forensics is applied to expose the vulnerability of forensics tools. Consequently, forensics researchers could develop forensics tools against possible new attacks. This dissertation investigation demonstrates two image forensics methods based on convolutional neural network (CNN) and two image anti-forensics methods based on generative adversarial network (GAN).

Detecting unsharp masking (USM) sharpened image is the first study in this dissertation. A CNN architecture comprises four convolutional layers and a classification module is proposed to discriminate sharpened images and unsharpened images. The results …

Land Cover Image Segmentation Based On Individual Class Binary Segmentation, Sathyanarayanan Somasunder

Theses

Remote sensing techniques have been developed over the past decades to acquire data without being in contact of the target object or data source. Their application on land-cover image segmentation has attracted significant attention in recent years. With the help of satellites, scientists and researchers can collect and store high resolution image data that can be further processed, segmented, and classified. However, these research results have not yet been synthesized to provide coherent guidance on the effect of variant land-cover segmentation processes. In this paper, we present a novel model that augments segmentation using smaller networks to segment individual classes. …

Rm-Net: Rasterizing Markov Signals To Images For Deep Learning, Kajal Gupta

Theses

Statistical machine learning approaches are quite famous for processing Markov signal data. They can model unobserved states and learn certain characteristics particular to a signal with good accuracy. However, with the advent of Deep learning the novice ways of solving a problem has shifted towards this more sophisticated algorithm, which is much better, powerful and more accurate. Specifically, Convolutional Neural Nets (CNN) have shown many promising results on images and videos. Here we illustrate how CNN can be applied to a 1D numeric signal using signal rasterization technique. We start by rasterizing a 1D numeric Markov signal into an image …

Automating Text Encapsulation Using Deep Learning, Anket Sah

Master's Projects

Data is an important aspect in any form be it communication, reviews, news articles, social media data, machine or real-time data. With the emergence of Covid-19, a pandemic seen like no other in recent times, information is being poured in from all directions on the internet. At times it is overwhelming to determine which data to read and follow. Another crucial aspect is separating factual data from distorted data that is being circulated widely. The title or short description of this data can play a key role. Many times, these descriptions can deceive a user with unwanted information. The user …

Improving Treatment Of Local Liver Ablation Therapy With Deep Learning And Biomechanical Modeling, Brian Anderson, Kristy Brock, Laurence Court, Carlos Eduardo Cardenas, Erik Cressman, Ankit Patel

Dissertations and Theses (Open Access)

In the United States, colorectal cancer is the third most diagnosed cancer, and 60-70% of patients will develop liver metastasis. While surgical liver resection of metastasis is the standard of care for treatment with curative intent, it is only avai lable to about 20% of patients. For patients who are not surgical candidates, local percutaneous ablation therapy (PTA) has been shown to have a similar 5-year overall survival rate. However, PTA can be a challenging procedure, largely due to spatial uncertainties in the localization of the ablation probe, and in measuring the delivered ablation margin.

For this work, we hypothesized …

A Deep Learning-Based Automatic Object Detection Method For Autonomous Driving Ships, Ojonoka Erika Atawodi

Master's Theses

An important feature of an Autonomous Surface Vehicles (ASV) is its capability of automatic object detection to avoid collisions, obstacles and navigate on their own.

Deep learning has made some significant headway in solving fundamental challenges associated with object detection and computer vision. With tremendous demand and advancement in the technologies associated with ASVs, a growing interest in applying deep learning techniques in handling challenges pertaining to autonomous ship driving has substantially increased over the years.

In this thesis, we study, design, and implement an object recognition framework that detects and recognizes objects found in the sea. We first curated …

Unveiling The Mystery Of Api Evolution In Deep Learning Frameworks: A Case Study Of Tensorflow 2, Zejun Zhang, Yanming Yang, Xin Xia, David Lo, Xiaoxue Ren, John C. Grundy

Research Collection School Of Computing and Information Systems

API developers have been working hard to evolve APIs to provide more simple, powerful, and robust API libraries. Although API evolution has been studied for multiple domains, such as Web and Android development, API evolution for deep learning frameworks has not yet been studied. It is not very clear how and why APIs evolve in deep learning frameworks, and yet these are being more and more heavily used in industry. To fill this gap, we conduct a large-scale and in-depth study on the API evolution of Tensorflow 2, which is currently the most popular deep learning framework. We first extract …

The Whole Is Greater Than Its Parts: Ensembling Improves Protein Contact Prediction, Wendy M. Billings, Connor J. Morris, Dennis Della Corte

Faculty Publications

The prediction of amino acid contacts from protein sequence is an important problem, as protein contacts are a vital step towards the prediction of folded protein structures. We propose that a powerful concept from deep learning, called ensembling, can increase the accuracy of protein contact predictions by combining the outputs of different neural network models. We show that ensembling the predictions made by different groups at the recent Critical Assessment of Protein Structure Prediction (CASP13) outperforms all individual groups. Further, we show that contacts derived from the distance predictions of three additional deep neural networks—AlphaFold, trRosetta, and ProSPr—can be substantially …

Network Traffic Anomaly Detection Method For Imbalanced Data, Shuqin Dong, Bin Zhang

Journal of System Simulation

Abstract: Aiming at the poor detection performances caused by the low feature extraction accuracy of rare traffic attacks from scarce samples, a network traffic anomaly detection method for imbalanced data is proposed. A traffic anomaly detection model is designed, in which the traffic features in different feature spaces are learned by alternating activation functions, architectures, corrupted rates and dropout rates of stacked denoising autoencoder (SDA), and the low accuracy in extracting features of rare traffic attacks in a single space is solved. A batch normalization algorithm is designed, and the Adam algorithm is adopted to train parameters of …

Wider Vision: Enriching Convolutional Neural Networks Via Alignment To External Knowledge Bases, Xuehao Liu, Sarah Jane Delany, Susan Mckeever

Conference papers

Deep learning models suffer from opaqueness. For Convolutional Neural Networks (CNNs), current research strategies for explaining models focus on the target classes within the associated training dataset. As a result, the understanding of hidden feature map activations is limited by the discriminative knowledge gleaned during training. The aim of our work is to explain and expand CNNs models via the mirroring or alignment of the network to an external knowledge base. This will allow us to give a semantic context or label for each visual feature. Using the resultant aligned embedding space, we can match CNN feature activations to nodes …

Wg2An: Synthetic Wound Image Generation Using Generative Adversarial Network, Salih Sarp, Murat Kuzlu, Emmanuel Wilson, Ozgur Guler

Engineering Technology Faculty Publications

In part due to its ability to mimic any data distribution, Generative Adversarial Network (GAN) algorithms have been successfully applied to many applications, such as data augmentation, text-to-image translation, image-to-image translation, and image inpainting. Learning from data without crafting loss functions for each application provides broader applicability of the GAN algorithm. Medical image synthesis is also another field that the GAN algorithm has great potential to assist clinician training. This paper proposes a synthetic wound image generation model based on GAN architecture to increase the quality of clinical training. The proposed model is trained on chronic wound datasets with various …

Addressing Artificial Intelligence Bias In Retinal Diagnostics, Philippe Burlina, Neil Joshi, William Paul, Katia D Pacheco, Neil M Bressler

Faculty, Staff and Students Publications

PURPOSE: This study evaluated generative methods to potentially mitigate artificial intelligence (AI) bias when diagnosing diabetic retinopathy (DR) resulting from training data imbalance or domain generalization, which occurs when deep learning systems (DLSs) face concepts at test/inference time they were not initially trained on.

METHODS: The public domain Kaggle EyePACS dataset (88,692 fundi and 44,346 individuals, originally diverse for ethnicity) was modified by adding clinician-annotated labels and constructing an artificial scenario of data imbalance and domain generalization by disallowing training (but not testing) exemplars for images of retinas with DR warranting referral (DR-referable) from darker-skin individuals, who presumably have greater …

Deep Learning For Multi-Tissue Cancer Classification Of Gene Expressions, Tarek Khorshed

Theses and Dissertations

We contribute in saving the lives of cancer patients through early detection and diagnosis, since one of the major challenges in cancer treatment is that patients are diagnosed at very late stages when appropriate medical interventions become less effective and full curative treatment is no longer achievable. Cancer classification using gene expressions is extremely challenging given the complexity and high dimensionality of the data. Current classification methods typically rely on samples collected from a single tissue type and perform a prerequisite of gene feature selection to avoid processing the full set of genes. These methods fall short in taking advantage …

Multi-Modal Classification Using Images And Text, Stuart J. Miller, Justin Howard, Paul Adams, Mel Schwan, Robert Slater

SMU Data Science Review

This paper proposes a method for the integration of natural language understanding in image classification to improve classification accuracy by making use of associated metadata. Traditionally, only image features have been used in the classification process; however, metadata accompanies images from many sources. This study implemented a multi-modal image classification model that combines convolutional methods with natural language understanding of descriptions, titles, and tags to improve image classification. The novelty of this approach was to learn from additional external features associated with the images using natural language understanding with transfer learning. It was found that the combination of ResNet-50 image …

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 …