Digital Commons Network^™

Learning Prototypes For Multiple Instance Learning, Özgür Emre Si̇vri̇kaya, Mert Yüksekgönül, Mustafa Gökçe Baydoğan

Turkish Journal of Electrical Engineering and Computer Sciences

Multiple instance learning (MIL) is a weakly supervised learning method that works on the labeled bag of instances data. A prototypical network is a popular embedding approach in MIL. They overcome the common problems that other MIL approaches may have to deal with including dimensionality, loss of instance-level information, and complexity. They demonstrate competitive performance in classification. This work proposes a simple model that provides a permutation invariant prototype generator from a given MIL data set. We aim to find out prototypes in the feature space to map the collection of instances (i.e. bags) to a distance feature space and …

Voice Impersonation For Thai Speech Using Cyclegan Over Prosody, Chatri Chuanngulueam

Chulalongkorn University Theses and Dissertations (Chula ETD)

No abstract provided.

A Two-Level Information Modelling Translation Methodology And Framework To Achieve Semantic Interoperability In Constrained Geoobservational Sensor Systems, Paul Stacey

Doctoral

As geographical observational data capture, storage and sharing technologies such as in situ remote monitoring systems and spatial data infrastructures evolve, the vision of a Digital Earth, first articulated by Al Gore in 1998 is getting ever closer. However, there are still many challenges and open research questions. For example, data quality, provenance and heterogeneity remain an issue due to the complexity of geo-spatial data and information representation.

Observational data are often inadequately semantically enriched by geo-observational information systems or spatial data infrastructures and so they often do not fully capture the true meaning of the associated datasets. Furthermore, data …

Secure Automatic Speaker Verification Systems, Muteb Aljasem

Wayne State University Dissertations

The growing number of voice-enabled devices and applications consider automatic speaker verification (ASV) a fundamental component. However, maximum outreach for ASV in critical domains e.g., financial services and health care, is not possible unless we overcome security breaches caused by voice cloning, and replayed audios collectively known as the spoofing attacks. The audio spoofing attacks over ASV systems on one hand strictly limit the usability of voice-enabled applications; and on the other hand, the counterfeiter also remains untraceable. Therefore, to overcome these vulnerabilities, a secure ASV (SASV) system is presented in this dissertation. The proposed SASV system is based on …

Evaluating The Performance Of Using Speaker Diarization For Speech Separation Of In-Person Role-Play Dialogues, Raveendra Medaramitta

Browse all Theses and Dissertations

Development of professional communication skills, such as motivational interviewing, often requires experiential learning through expert instructor-guided role-plays between the trainee and a standard patient/actor. Due to the growing demand for such skills in practices, e.g., for health care providers in the management of mental health challenges, chronic conditions, substance misuse disorders, etc., there is an urgent need to improve the efficacy and scalability of such role-play based experiential learning, which are often bottlenecked by the time-consuming performance assessment process. WSU is developing ReadMI (Real-time Assessment of Dialogue in Motivational Interviewing) to address this challenge, a mobile AI solution aiming to …

Evaluating The Performance Of Using Speaker Diarization For Speech Separation Of In-Person Role-Play Dialogues, Raveendra Medaramitta

Browse all Theses and Dissertations

Development of professional communication skills, such as motivational interviewing, often requires experiential learning through expert instructor-guided role-plays between the trainee and a standard patient/actor. Due to the growing demand for such skills in practices, e.g., for health care providers in the management of mental health challenges, chronic conditions, substance misuse disorders, etc., there is an urgent need to improve the efficacy and scalability of such role-play based experiential learning, which are often bottlenecked by the time-consuming performance assessment process. WSU is developing ReadMI (Real-time Assessment of Dialogue in Motivational Interviewing) to address this challenge, a mobile AI solution aiming to …

Automated Classification Of Bi-Rads In Textual Mammography Reports, Mostafa Boroumandzadeh, Elham Parvinnia

Turkish Journal of Electrical Engineering and Computer Sciences

The main purpose of this paper is to process key information in medical text records and also classifypatients, per different levels of breast imaging-reporting and data system (BI-RADS). The BI-RADS is a scheme for thestandardization of breast imaging reports. Therefore, medical text mining is employed to classify mammography reportssupported BI-RADS. In this research, a new method is proposed for automated BI-RADS classifications extraction fromtextual reports and improves the therapeutic procedures. At first, a mammography lexicon is employed for choosingkeywords from medical text reports. Word2vec and term frequency inverse document frequency (TFIDF) techniques areused for extracting features, finally, they are combined …

A Fuzzy Expert System For Predicting The Mortality Of Covid'19, Monika Mangla, Nonita Sharma, Poonam Mittal

Turkish Journal of Electrical Engineering and Computer Sciences

The COVID-19 pandemic has had a widespread impact on health and economy across the globe. It is leading to a huge number of deaths per day. Few researchers have been attracted to analyzing the mortality rate of COVID-19 from various perspectives. During the research, it has become evident that these fatalities are not only caused by COVID19, but they are also affected by some other factors. The authors of this paper aim to encompass three important types of factors viz. risk factors, clinical factors, and miscellaneous factors that influence the mortality of COVID-19. This manuscript presents a rule-based model under …

Aviation English Assessment And Training, Robert Fowler, Elizabeth Mathews, Jena Lynch, Jennifer Roberts M.A.

Publications

Due to a significant global increase in demand for air travel, there has been a corresponding increase in demand for ab initio flight training. Thousands of international flight students seek admission to collegiate aviation programs in the United States and Canada every year. These international flight students come to the United States and Canada because flight training is nonexistent in their native countries. In fact, flight training in most of these countries is impossible due to airspace restrictions and onerous regulations. If there is flight training available in these countries, the cost is usually prohibitive compared to the cost in …

New Features For Speech Processing Standard Pronunciation Classification, Mutian Zhuang

Graduate Research Theses & Dissertations

Pronunciation is essential in language learning. Standard English speakers could pronounce in the right way, because their pronunciation procedure is correct. The aim of this research is to see how an artificial intelligence program can be developed to provide feedback to people while they learn to generate accurate phoneme /r/. The audio signal will first be detected, after which features of audio signal will be extracted, and finally, classification will be performed. This suggested algorithm and method may be applied to other related topics. This thesis proposes new features and uses them to classify standard English speakers and non-standard English …

Nlp Is Not Enough - Contextualization Of User Input In Chatbots, Nathan Dolbir, Triyasha Dastidar, Kaushik Roy

Publications

AI chatbots have made vast strides in technology improvement in recent years and are already operational in many industries. Advanced Natural Language Processing techniques, based on deep networks, efficiently process user requests to carry out their functions. As chatbots gain traction, their applicability in healthcare is an attractive proposition due to the reduced economic and people costs of an overburdened system. However, healthcare bots require safe and medically accurate information capture, which deep networks aren’t yet capable of due to user text and speech variations. Knowledge in symbolic structures is more suited for accurate reasoning but cannot handle natural language …

Semantics Of The Black-Box: Can Knowledge Graphs Help Make Deep Learning Systems More Interpretable And Explainable?, Manas Gaur, Keyur Faldu, Amit Sheth

Publications

The recent series of innovations in deep learning (DL) have shown enormous potential to impact individuals and society, both positively and negatively. The DL models utilizing massive computing power and enormous datasets have significantly outperformed prior historical benchmarks on increasingly difficult, well-defined research tasks across technology domains such as computer vision, natural language processing, signal processing, and human-computer interactions. However, the Black-Box nature of DL models and their over-reliance on massive amounts of data condensed into labels and dense representations poses challenges for interpretability and explainability of the system. Furthermore, DLs have not yet been proven in their ability to …

Incorporating Context Into Non-Autoregressive Model Using Contextualized Ctc For Sequence Labelling, Burin Naowarat

Chulalongkorn University Theses and Dissertations (Chula ETD)

Connectionist Temporal Classification (CTC) loss has become widely used in sequence modeling tasks such as Automatic Speech Recognition (ASR) and Handwritten Text Recognition (HTR) due to its ease of use. CTC itself has no architecture constraints, but it is commonly used with recurrent models that predict letters based on histories in order to relax the conditional independent assumption. However, recent sequence models that incorporate CTC loss have been focusing on speed by removing recurrent structures, hence losing important context information. This thesis presents Contextualized Connectionist Temporal Classification (CCTC) loss, which induces prediction dependencies in non-recurrent and non-autoregressive neural networks for …

Fuzzy Ahp For Supplier Selection: A Case Study In An Electronic Component Manufacturer, Chia Ken Tsai

Chulalongkorn University Theses and Dissertations (Chula ETD)

Supplier selection has become an essential effect on the entire electronic supply chain network on performance. The case study company produces a nano sim-card connector which four primary raw materials are processed into four primary parts (Plastic, Nickel, Phosphor bronze, and Stainless steel). Nevertheless, the case study company faces a penalty and order reduction because of the quality issue. Although an appraisal record from the case study company is able to select a proper raw material supplier, the cost becomes the priority when the candidate suppliers are categorized as the same level, leading to increasing potential risks, e.g., a penalty, …

Using Deep Learning-Based Framework For Child Speech Emotion Recognition, Gerald N. Onwujekwe

Theses and Dissertations

Biological languages of the body through which human emotion can be detected abound including heart rate, facial expressions, movement of the eyelids and dilation of the eyes, body postures, skin conductance, and even the speech we make. Speech emotion recognition research started some three decades ago, and the popular Interspeech Emotion Challenge has helped to propagate this research area. However, most speech recognition research is focused on adults and there is very little research on child speech. This dissertation is a description of the development and evaluation of a child speech emotion recognition framework. The higher-level components of the framework …

A Case Study On Player Selection And Team Formation In Football With Machinelearning, Di̇dem Abi̇di̇n

Turkish Journal of Electrical Engineering and Computer Sciences

Machine learning has been widely used in different domains to extract information from raw data. Sports is one of the popular domains for researchers to work on recently. Although score prediction for matches is the most preferred application area for artificial intelligence, player selection, and team formation is also an application area worth working on. There are some studies in the literature about player selection and team formation which are examined in this study. The study has two important contributions: First one is to apply seven different machine learning algorithms on our dataset to find the best player combination for …

Hybrid Mamdani Fuzzy Rules And Convolutional Neural Networks For Analysis And Identification Of Animal Images, Hind R. Mohammed, Zahir M. Hussain

Research outputs 2014 to 2021

Accurate, fast, and automatic detection and classification of animal images is challenging, but it is much needed for many real-life applications. This paper presents a hybrid model of Mamdani Type-2 fuzzy rules and convolutional neural networks (CNNs) applied to identify and distinguish various animals using different datasets consisting of about 27,307 images. The proposed system utilizes fuzzy rules to detect the image and then apply the CNN model for the object’s predicate category. The CNN model was trained and tested based on more than 21,846 pictures of animals. The experiments’ results of the proposed method offered high speed and efficiency, …

Exposure Assessment Of Emerging Contaminants: Rapid Screening And Modeling Of Plant Uptake, Majid Bagheri

Doctoral Dissertations

"With the advent of new chemicals and their increasing uses in every aspect of our life, considerable number of emerging contaminants are introduced to environment yearly. Emerging contaminants in forms of pharmaceuticals, detergents, biosolids, and reclaimed wastewater can cross plant roots and translocate to various parts of the plants. Long-term human exposure to emerging contaminants through food consumption is assumed to be a pathway of interest. Thus, uptake and translocation of emerging contaminants in plants are important for the assessment of health risks associated with human exposure to emerging contaminants. To have a better understanding over fate of emerging contaminants …

Detecting Interlocutor Confusion In Situated Human-Avatar Dialogue: A Pilot Study, Na Li, John D. Kelleher, Robert J. Ross

Conference papers

In order to enhance levels of engagement with conversational systems, our long term research goal seeks to monitor the confusion state of a user and adapt dialogue policies in response to such user confusion states. To this end, in this paper, we present our initial research centred on a user-avatar dialogue scenario that we have developed to study the manifestation of confusion and in the long term its mitigation. We present a new definition of confusion that is particularly tailored to the requirements of intelligent conversational system development for task-oriented dialogue. We also present the details of our Wizard-of-Oz based …

Trust-Empowered, Iot-Driven Legitimate Data Offloading, Zakaria Maamar, Noura Faci, Fadwa Yahya, Ejub Kajan

All Works

In an IoT environment deployed on top of fog and/or cloud nodes, offloading data between nodes is a common practice that aims at lessening the burden on these nodes and hence, meeting some real-time processing requirements. Existing initiatives put emphasis on “when to offload” and “where to offload” using criteria like resource constraint, load balancing, and data safety during transfer. However, there is limited emphasis on the trustworthiness of those nodes that will accept the offloaded data putting these data at risk of misuse. To address this limited emphasis, this paper advocates for trust as a decision criterion for identifying …

Parents' Experiences Of A Language-Focused Home Visiting Scheme In Ireland, Aisling Ni Dhiorbhain Dr, Maire Mhic Mhathuna, Padraig Ó Duibhir Dr

Articles

This article reports parents’ experiences of the Tús Maith (Good Start) home visiting scheme in South-West Ireland. The goal of Tús Maith is to support parents who wish to speak Irish to their children at home in the Kerry Gaeltacht, an Irish-speaking heartland area. Home visitors spend an hour a week, over a period of six weeks, interacting with children and parents with varying levels of competency in Irish. Home visitors who are native speakers of Irish, offer individualised guidance on how to promote the use of Irish as a home language, while encouraging families to engage in activities …

Mental Health And The Covid-19 Pandemic: Analysis Of Twitter Discourse, Omar El-Gayar, Abdullah Wahbeh, Tareq Nasralah, Ahmed El Noshokaty, Mohammad A. Al-Ramahi

Computer Information Systems Faculty Publications (Archived)

This study analyzed Twitter discourse to understand the association of the COVID-19 pandemic with mental health. The study compared tweets’ volume over time, tweets’ volume per mental health category, emotions, and the top hashtags on mental health before and after November 2019, the month on which the first COVID-19 case was reported. We analyzed a total of 273 million English tweets on mental health collected from 56 million unique users. Results and analysis showed a significant shift in trend for the volume of tweets on mental health over time. There was also a notable increase in the volume of tweets …

Evaluating Social Media As A Medium Of Private Communication Through Steganographic Images, Kendall Weldon Coles

Graduate Theses, Dissertations, and Problem Reports (ETD)

Social media is a vastly used communication tool with billions of users worldwide. These social networks provide users the ability to share their ideas and thoughts through the messages, videos, and images that they post. The images that are shared can possibly be embedded with private messages through the use of a steganographic tool. The messages are embedded in a fashion that doesn’t change the visual appearance of an image. This allows for these types of images to hide in plain sight, which creates the possibility of someone communicating privately in a public social media setting. This project proved how …

Deep Models For Improving The Performance And Reliability Of Person Recognition, Sobhan Soleymani

Graduate Theses, Dissertations, and Problem Reports (ETD)

Deep models have provided high accuracy for different applications such as person recognition, image segmentation, image captioning, scene description, and action recognition. In this dissertation, we study the deep learning models and their application in improving the performance and reliability of person recognition. This dissertation focuses on five aspects of person recognition: (1) multimodal person recognition, (2) quality-aware multi-sample person recognition, (3) text-independent speaker verification, (4) adversarial iris examples, and (5) morphed face images. First, we discuss the application of multimodal networks consisting of face, iris, fingerprint, and speech modalities in person recognition. We propose multi-stream convolutional neural network architectures …

Interactive Modules For Flight Training: A Review, Stephanie G. Fussell, Robert Thomas

Journal of Aviation/Aerospace Education & Research

The use of animation and video, combined with auditory and text components, to teach complex subjects can be beneficial to deepen understanding. A review of the literature was conducted to better understand how interactive learning modules can give flight students tools to enhance the learning process outside of traditional instruction. Instructional theories and educational environments were considered through the lens of designing course content for flight students. Several research questions were asked to guide the course of the research, all focused on how students learn a complex subject matter using interactive material. Guidelines and principles are discussed to ensure the …

Bond Graph Modeling Of Critical Infrastructures For Cyber-Physical Security Implementation, Michele Jane White

Masters Theses

“In developed societies, there exists infrastructure vital to everyday life. This includes water and power systems. Technology is quickly evolving and being implemented on these utilities. This technology can range from smart metering in neighborhoods to volume sensors in local waste water treatment facilities. When networking, sensing, monitoring, or control devices are integrated with infrastructure it is considered a cyber physical system, or CPS. When information about an important physical system is connected to the virtual world, it is opened up to security risks. Cyber security can be provided to the CPS by monitoring the physical state of the system …

Kendali Fuzzy Logic - Interleaved Boost Converter Pada Aplikasi Motor Dc, Mentari Putri Jati, Gamar Basuki, Hasnira Hasnira

Elinvo (Electronics, Informatics, and Vocational Education)

Karakteristik motor DC menjadi hal mutlak yang perlu diketahui dalam proses pengendalian kecepatan penggerak elektrik. Konverter pengemudian elektrik yang sedang berkembang yaitu Interleaved Boost Converter. Pengendali fuzzy logic merupakan kendali yang efektif diterapkan banyak sistem linier maupun nonlinier dengan waktu operasi yang cepat. Hasil penerapan aplikasi motor DC berupa penyiram kecambah dapat berfungsi dengan baik pada variasi nilai setpoint kecepatan yang mana perubahan nilai tegangan keluaran berbanding lurus dengan perubahan kecepatan motor DC (30 Volt sd. 60 Volt ‰ˆ 800 Rpm sd 1500 Rpm). Hal ini dapat membantu para petani kecambah dalam melakukan penyiraman biji kecambah secara otomatis seiring …

Model-Based Approach For Product Requirement Representation And Generation In Product Lifecycle Management, Omer Yaman

Dissertations - ALL

The requirement specification is an official documentation activity, which is a collection of certain information to specify the product and its life-cycle activities in terms of functions, features, performance, constraints, production, maintenance, disposal process, etc. It contains mainly two phases; product requirement generation and representation. Appropriate criteria for the product design and further life-cycle activities are determined based on the requirement specification as well as the interrelations of product requirements with other life-cycle information such as; materials, manufacturing, working environments, finance, and regulations. The determination of these criteria is normally error-prone. It is difficult to identify and maintain the completeness …

Using Eye-Gaze To Evaluate Neural Attention, Shahansha Salim

Master’s Dissertations

The ability to selectively concentrate on areas of interest while ignoring the rest is termed as attention in human beings. This ability has played a key role in survival as well as information processing. Neural Attention is said to be an effort to bring similar action of selectively concentrating areas of relevance in deep neural networks. This simple yet powerful concept has attracted a lot of research in recent years, yielding breakthrough results in Natural Language Processing (NLP) problems and main stream Computer Vision problems such as Image Caption Generation, Neural Machine Translation (NMT), Visual Question Answering (VQA), Action Recognition, …

Technology Criticism And Data Literacy: The Case For An Augmented Understanding Of Media Literacy, Thomas Knaus

Journal of Media Literacy Education

Reviewing the history of media literacy education might help us to identify how creating media as an approach can contribute to fostering knowledge, understanding technical issues, and to establishing a critical attitude towards technology and data. In a society where digital devices and services are omnipresent and decisions are increasingly based on data, critical analysis must penetrate beyond the “outer shell” of machines – their interfaces – through the technology itself, and the data, and algorithms, which make these devices and services function. Because technology and data constitute the basis of all communication and collaboration, media literate individuals …