Computer Sciences Commons^™

Learning To Fuse Asymmetric Feature Maps In Siamese Trackers, Wencheng Han, Xingping Dong, Fahad Shahbaz Khan, Ling Shao, Jianbing Shen

Computer Vision Faculty Publications

Recently, Siamese-based trackers have achieved promising performance in visual tracking. Most recent Siamese-based trackers typically employ a depth-wise cross-correlation (DW-XCorr) to obtain multi-channel correlation information from the two feature maps (target and search region). However, DW-XCorr has several limitations within Siamese-based tracking: it can easily be fooled by distractors, has fewer activated channels and provides weak discrimination of object boundaries. Further, DW-XCorr is a handcrafted parameter-free module and cannot fully benefit from offline learning on large-scale data. We propose a learnable module, called the asymmetric convolution (ACM), which learns to better capture the semantic correlation information in offline training on …

Script For Estimating Error And Bias In Offline Evaluation Results, Mucun Tian, Michael D. Ekstrand

Computer Science Faculty Scripts and Data

This publication contains scripts to reproduce the paper “Estimating Error and Bias in Offline Evaluation Results” by Muncun Tian and Michael D. Ekstrand in Proceedings of the 2020 Conference on Human Information Interaction and Retrieval (CHIIR '20).

Script For Exploring Author Gender In Book Rating And Recommendation, Michael D. Ekstrand, Daniel Kluver

Computer Science Faculty Scripts and Data

This publication contains scripts to reproduce the paper:

Ekstrand, M.D. and Kluver, D. (2021). Exploring Author Gender in Book Rating and Recommendation. User Modeling and User-Adapted Interaction. https://doi.org/10.1007/s11257-020-09284-2

*Date reflected refers to the publisher's online early release date.

Analysis Of Contextual Emotions Using Multimodal Data, Saurabh Hinduja

USF Tampa Graduate Theses and Dissertations

Affective computing builds and evaluates systems that can recognize, interpret, and simulate human emotion. It is an interdisciplinary field, which includes computer science, psychology, and many others. For years, human emotion has been studied in psychology but recently has become a prominent field in computer science. Largely, the field of affective computing has been focused on analyzing static facial expressions to recognize human emotions, without taking bias (e.g. gender, data bias), context, or temporal information into account. Psychology has shown the difficulty of analyzing emotions without incorporating this type of information. In this dissertation, we have proposed new approaches to …

Exhpd: Exploiting Human, Physical, And Driving Behaviors To Detect Vehicle Cyber Attacks, Qian Chen, Paul Romanowich, Jorge Castillo, Krishna Chandra Roy, Gustavo Chavez, Shouhuai Xu

Informatics and Engineering Systems Faculty Publications

As increasingly more vehicles are connected to the Internet, cyber attacks against vehicles are becoming a real threat with devastating consequences. This highlights the importance of detecting vehicle cyber attacks before fatal accidents occur. One natural method for tackling this problem is to adapt existing approaches for detecting attacks in enterprize networks, but which has achieved limited success. In this article, we propose a new approach to treat vehicles as cyber-physical-human systems, leading to a novel framework called exploiting human, physical and driving behaviors to detect vehicle cyber attacks (ExHPD). The framework has four detectors: 1) a human detector; 2) …

Deep Gaussian Processes For Few-Shot Segmentation, Joakim Johnander, Johan Edstedt, Martin Danelljan, Michael Felsberg, Fahad Shahbaz Khan

Computer Vision Faculty Publications

Few-shot segmentation is a challenging task, requiring the extraction of a generalizable representation from only a few annotated samples, in order to segment novel query images. A common approach is to model each class with a single prototype. While conceptually simple, these methods suffer when the target appearance distribution is multi-modal or not linearly separable in feature space. To tackle this issue, we propose a few-shot learner formulation based on Gaussian process (GP) regression. Through the expressivity of the GP, our approach is capable of modeling complex appearance distributions in the deep feature space. The GP provides a principled way …

Efficient Post-Quantum And Compact Cryptographic Constructions For The Internet Of Things, Rouzbeh Behnia

USF Tampa Graduate Theses and Dissertations

IoT systems often rely on low-end devices to send measurements to other parties and depending on the setting, unauthorized alteration and/or privacy violation of these measures can have catastrophic consequences (e.g., embedded medical sensors). Therefore, providing efficient authentication, integrity, and confidentiality in these settings is vital. While conventional cryptographic measures (e.g., ECDSA) can be used to meet these security requirements, despite their elegant design, they are often too computationally expensive for low-end devices. This is further exacerbated when security against quantum computers is taken into the account.

In this dissertation, we propose a series of new efficient conventional and post-quantum …

Improving Memory Forensics Through Emulation And Program Analysis, Ryan Dominick Maggio

LSU Doctoral Dissertations

Memory forensics is an important tool in the hands of investigators. However, determining if a computer is infected with malicious software is time consuming, even for experts. Tasks that require manual reverse engineering of code or data structures create a significant bottleneck in the investigative workflow. Through the application of emulation software and symbolic execution, these strains have been greatly lessened, allowing for faster and more thorough investigation. Furthermore, these efforts have reduced the barrier for forensic investigation, so that reasonable conclusions can be drawn even by non-expert investigators. While previously Volatility had allowed for the detection of malicious hooks …

Mathematical Modeling Of The Candida Albicans Yeast To Hyphal Transition Reveals Novel Control Strategies, David J. Wooten, Jorge Gómez Tejeda Zañudo, David Murrugarra, Austin M. Perry, Anna Dongari-Bagtzoglou, Reinhard Laubenbacher, Clarissa J. Nobile, Réka Albert

Mathematics Faculty Publications

Candida albicans, an opportunistic fungal pathogen, is a significant cause of human infections, particularly in immunocompromised individuals. Phenotypic plasticity between two morphological phenotypes, yeast and hyphae, is a key mechanism by which C. albicans can thrive in many microenvironments and cause disease in the host. Understanding the decision points and key driver genes controlling this important transition and how these genes respond to different environmental signals is critical to understanding how C. albicans causes infections in the host. Here we build and analyze a Boolean dynamical model of the C. albicans yeast to hyphal transition, integrating …

Integrated Cyberattack Detection And Handling For Nonlinear Systems With Evolving Process Dynamics Under Lyapunov-Based Economic Model Predictive Control, Keshav Kasturi Rangan, Henrique Oyama, Helen Durand

Chemical Engineering and Materials Science Faculty Research Publications

Safety-critical processes are becoming increasingly automated and connected. While automation can increase effciency, it brings new challenges associated with guaranteeing safety in the presence of uncertainty especially in the presence of control system cyberattacks. One of the challenges for developing control strategies with guaranteed safety and cybersecurity properties under suffcient conditions is the development of appropriate detection strategies that work with control laws to prevent undetected attacks that have immediate closed-loop stability consequences. Achieving this, in the presence of uncertainty brought about by plant/model mismatch and process dynamics that can change with time, requires a fundamental understanding of the characteristics …

Network-Based Analysis Of Early Pandemic Mitigation Strategies: Solutions, And Future Directions, Pegah Hozhabrierdi, Raymond Zhu, Maduakolam Onyewu, Sucheta Soundarajan

Northeast Journal of Complex Systems (NEJCS)

Despite the large amount of literature on mitigation strategies for pandemic spread, in practice, we are still limited by naive strategies, such as lockdowns, that are not effective in controlling the spread of the disease in long term. One major reason behind adopting basic strategies in real-world settings is that, in the early stages of a pandemic, we lack knowledge of the behavior of a disease, and so cannot tailor a more sophisticated response. In this study, we design different mitigation strategies for early stages of a pandemic and perform a comprehensive analysis among them. We then propose a novel …

On Generating Transferable Targeted Perturbations, Muzammal Naseer, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, Fatih Porikli

Computer Vision Faculty Publications

While the untargeted black-box transferability of adversarial perturbations has been extensively studied before, changing an unseen model's decisions to a specific 'targeted' class remains a challenging feat. In this paper, we propose a new generative approach for highly transferable targeted perturbations (TTP). We note that the existing methods are less suitable for this task due to their reliance on class-boundary information that changes from one model to another, thus reducing transferability. In contrast, our approach matches the perturbed image 'distribution' with that of the target class, leading to high targeted transferability rates. To this end, we propose a new objective …

Optimizing Networking Topologies With Shortest Path Algorithms, Jordan Sahs

UNO Student Research and Creative Activity Fair

Communication networks tend to contain redundant devices and mediums of transmission, thus the need to locate, document, and optimize networks is increasingly becoming necessary. However, many people do not know where to start the optimization progress. What is network topology? What is this “Shortest Path Problem”, and how can it be used to better my network? These questions are presented, taught, and answered within this paper. To supplement the reader’s understanding there are thirty-eight figures in the paper that are used to help convey and compartmentalize the learning process needed to grasp the materials presented in the ending sections.

In …

What If Keys Are Leaked? Towards Practical And Secure Re-Encryption In Deduplication-Based Cloud Storage, Weijing You, Lei Lei, Bo Chen, Limin Liu

Michigan Tech Publications, Part 1

By only storing a unique copy of duplicate data possessed by different data owners, deduplication can significantly reduce storage cost, and hence is used broadly in public clouds. When combining with confidentiality, deduplication will become problematic as encryption performed by different data owners may differentiate identical data which may then become not deduplicable. The Message-Locked Encryption (MLE) is thus utilized to derive the same encryption key for the identical data, by which the encrypted data are still deduplicable after being encrypted by different data owners. As keys may be leaked over time, re-encrypting outsourced data is of paramount importance to …

Building And Using Digital Libraries For Etds, Edward A. Fox

The Journal of Electronic Theses and Dissertations

Despite the high value of electronic theses and dissertations (ETDs), the global collection has seen limited use. To extend such use, a new approach to building digital libraries (DLs) is needed. Fortunately, recent decades have seen that a vast amount of “gray literature” has become available through a diverse set of institutional repositories as well as regional and national libraries and archives. Most of the works in those collections include ETDs and are often freely available in keeping with the open-access movement, but such access is limited by the services of supporting information systems. As explained through a set of …

Orthogonal Projection Loss, Kanchana Ranasinghe, Muzammal Naseer, Munawar Hayat, Salman Khan, Fahad Shahbaz Khan

Computer Vision Faculty Publications

Deep neural networks have achieved remarkable performance on a range of classification tasks, with softmax cross-entropy (CE) loss emerging as the de-facto objective function. The CE loss encourages features of a class to have a higher projection score on the true class-vector compared to the negative classes. However, this is a relative constraint and does not explicitly force different class features to be well-separated. Motivated by the observation that ground-truth class representations in CE loss are orthogonal (one-hot encoded vectors), we develop a novel loss function termed 'Orthogonal Projection Loss' (OPL) which imposes orthogonality in the feature space. OPL augments …

Automatic Detection Of Vehicles In Satellite Images For Economic Monitoring, Cole Hill

USF Tampa Graduate Theses and Dissertations

With the growing supply of satellites capturing images of the planet, governments andinvestors are looking for ways in which these new images may be used to determine which businesses are struggling and thriving. Recent works have shown that parking lot fill rates can provide valuable information about businesses’ earnings, however, the task of manually annotating the number of vehicles in a parking lot is expensive and time-consuming. Systems which can automate this process are therefore valuable as they are faster and cheaper than human labor. In this thesis, the problem of detection of small objects in large low-resolution images is …

“When They Say Weed Causes Depression, But It’S Your Fav Antidepressant”: Knowledge-Aware Attention Framework For Relationship Extraction, Shweta Yadav, Usha Lokala, Raminta Daniulaityte, Krishnaprasad Thirunarayan, Francois Lamy, Amit P. Sheth

Faculty Publications

With the increasing legalization of medical and recreational use of cannabis, more research is needed to understand the association between depression and consumer behavior related to cannabis consumption. Big social media data has potential to provide deeper insights about these associations to public health analysts. In this interdisciplinary study, we demonstrate the value of incorporating domain-specific knowledge in the learning process to identify the relationships between cannabis use and depression. We develop an end-to-end knowledge infused deep learning framework (Gated-K-BERT) that leverages the pre-trained BERT language representation model and domain-specific declarative knowledge source (Drug Abuse Ontology) to jointly extract entities …

Neuro-Symbolic Deductive Reasoning For Cross-Knowledge Graph Entailment, Monireh Ebrahimi, Md Kamruzzaman Sarker, Federico Bianchi, Ning Xie, Aaron Eberhart, Derek Doran, Hyeongsik Kim, Pascal Hitzler

Computer Science and Engineering Faculty Publications

A significant and recent development in neural-symbolic learning are deep neural networks that can reason over symbolic knowledge graphs (KGs). A particular task of interest is KG entailment, which is to infer the set of all facts that are a logical consequence of current and potential facts of a KG. Initial neural-symbolic systems that can deduce the entailment of a KG have been presented, but they are limited: current systems learn fact relations and entailment patterns specific to a particular KG and hence do not truly generalize, and must be retrained for each KG they are tasked with entailing. We …

Analysis Of System Performance Metrics Towards The Detection Of Cryptojacking In Iot Devices, Richard Matthews

Masters Theses & Doctoral Dissertations

This single-case mechanism study examined the effects of cryptojacking on Internet of Things (IoT) device performance metrics. Cryptojacking is a cyber-threat that involves stealing the computational resources of devices belonging to others to generate cryptocurrencies. The resources primarily include the processing cycles of devices and the additional electricity needed to power this additional load. The literature surveyed showed that cryptojacking has been gaining in popularity and is now one of the top cyberthreats. Cryptocurrencies offer anyone more freedom and anonymity than dealing with traditional financial institutions which make them especially attractive to cybercriminals. Other reasons for the increasing popularity of …

Analyzing The Effectiveness Of Legal Regulations And Social Consequences For Securing Data, Howard B. Goodman

Masters Theses & Doctoral Dissertations

There is a wide range of concerns and challenges related to stored data security – which range from privacy and management to operations readiness, These challenges span from financial to personal and public impact. With an abundance of regulations for the enforcement of data security and emerging requirements proposed every year, organizations cannot avoid the legal or social implications of inadequate data protection. Today, public spotlight and awareness are challenging organizations to enhance how data is protected more than at any other time. For this reason, organizations have made significant efforts to improve security.

When looking at precautions or changes, …

Survey On Quantum Circuit Compilation For Noisy Intermediate-Scale Quantum Computers: Artificial Intelligence To Heuristics, Janusz Kusyk, Samah Mohamed Saeed, Muharrem Umit Uyar

Publications and Research

Computationally expensive applications, including machine learning, chemical simulations, and financial modeling, are promising candidates for noisy intermediate scale quantum (NISQ) computers. In these problems, one important challenge is mapping a quantum circuit onto NISQ hardware while satisfying physical constraints of an underlying quantum architecture. Quantum circuit compilation (QCC) aims to generate feasible mappings such that a quantum circuit can be executed in a given hardware platform with acceptable confidence in outcomes. Physical constraints of a NISQ computer change frequently, requiring QCC process to be repeated often. When a circuit cannot directly be executed on a quantum hardware due to its …

Scite: The Next Generation Of Citations, Sean Rife, Domenic Rosati, Joshua M. Nicholson

Faculty & Staff Research and Creative Activity

Key points

• While the importance of citation context has long been recognized, simple citation counts remain as a crude measure of importance.
• Providing citation context should support the publication of careful science instead of headline‐grabbing and salami‐sliced non‐replicable studies.
• Machine learning has enabled the extraction of citation context for the first time, and made the classification of citation types at scale possible.

The Impact Of Twitter On The National Hockey League And Its Players, Benjamin Strauss

Honors Projects in Data Science

This study offers a new perspective on collecting and analyzing Twitter data surrounding the National Hockey League (NHL) to identify any trends or relationships between the data and overall performance during the 2021 abbreviated season. This paper provides and in-depth analysis by studying a sample of sixty of the top NHL players, specifically those who are typically top performers in the league, spanning over all thirty-one teams and all positions, this study was able to identify a deeper and broader perspective of what implications can be drawn from analyzing data from Twitter to both predict and reflect both individual player …

Hybrid Deep Learning Architecture To Forecast Maximum Load Duration Using Time-Of-Use Pricing Plans, Jinseok Kim, Babar Shah, Ki Il Kim

All Works

Load forecasting has received crucial research attention to reduce peak load and contribute to the stability of power grid using machine learning or deep learning models. Especially, we need the adequate model to forecast the maximum load duration based on time-of-use, which is the electricity usage fare policy in order to achieve the goals such as peak load reduction in a power grid. However, the existing single machine learning or deep learning forecasting cannot easily avoid overfitting. Moreover, a majority of the ensemble or hybrid models do not achieve optimal results for forecasting the maximum load duration based on time-of-use. …

Encryption And Decryption With A Raspberry Pi Device, Taylor Powell

Undergraduate Research Symposium

The functioning of our modern digital world relies heavily on the security of modern encryption algorithms and their resistance to systematic attempts to access secure information. For the 2020 Department of Computer Science’s Raspberry Pi Programming Competition, I decided to explore encryption and decryption techniques available to any user with some programming knowledge and a desire to secure information from unwanted access.

I developed a program which allows a user to select between three types of encryption algorithms: a Caesar Cipher, a Vigenère Cipher, and a Stream Cipher. I also gave the user the option to further secure their encrypted …

Trends Observation: Hot Research Field Of Information Technology From 2017 To 2020

Bulletin of Chinese Academy of Sciences (Chinese Version)

No abstract provided.

Qosa-Icn: An Information-Centric Approach To Qos In Vehicular Environments, Jessica Mccarthy, Saqib Rasool Chaudhry, Perumal Kuppuudaiyar, Radhika Loomba, Siobhan Clarke

Department of Computer Science Publications

Heterogeneous content-based traffic distribution motivates Information-Centric Networking (ICN), where content delivery is of primary interest as a prominent solution. However, current work does not address Quality of Service (QoS) provisioning for prioritized traffic, which is required for different applications and content types. This paper extends ICN with data delivery deadline awareness and shapes the forwarding decisions to ensure prioritized packet treatment. The proposed QoS Aware-ICN (QoSA-ICN) classifies requests' priority with their QoS requirements by codifying a QoSInfo object in interest/data packets. QoSA-ICN also extends the existing NDN transmission mode to a converged best route with multi-hop multi-route forwarding, to avoid …

Leveraging Natural Language Processing To Mine Issues On Twitter During The Covid-19 Pandemic, Ankita Agarwal, Preetham Salehundam, Swati Padhee, William Romine, Tanvi Wright State University - Main Campus

Computer Science and Engineering Faculty Publications

The recent global outbreak of the coronavirus disease (COVID-19) has spread to all corners of the globe. The international travel ban, panic buying, and the need for self-quarantine are among the many other social challenges brought about in this new era. Twitter platforms have been used in various public health studies to identify public opinion about an event at the local and global scale. To understand the public concerns and responses to the pandemic, a system that can leverage machine learning techniques to filter out irrelevant tweets and identify the important topics of discussion on social media platforms like Twitter …

Topic-Centric Unsupervised Multi-Document Summarization Of Scientific And News Articles, Amanuel Alambo, Cori Lohstroh, Erik Madaus, Swati Padhee, Brandy Foster, Tanvi Banerjee, Krishnaprasad Thirunarayan, Michael Raymer

Computer Science and Engineering Faculty Publications

Recent advances in natural language processing have enabled automation of a wide range of tasks, including machine translation, named entity recognition, and sentiment analysis. Automated summarization of documents, or groups of documents, however, has remained elusive, with many efforts limited to extraction of keywords, key phrases, or key sentences. Accurate abstractive summarization has yet to be achieved due to the inherent difficulty of the problem, and limited availability of training data. In this paper, we propose a topic-centric unsupervised multi-document summarization framework to generate extractive and abstractive summaries for groups of scientific articles across 20 Fields of Study (FoS) in …