Digital Commons Network^™

Hategan: Adversarial Generative-Based Data Augmentation For Hate Speech Detection, Rui Cao, Roy Ka-Wei Lee

SCIS Student Publications

Academia and industry have developed machine learning and natural language processing models to detect online hate speech automatically. However, most of these existing methods adopt a supervised approach that heavily depends on labeled datasets for training. This results in the methods’ poor detection performance of the hate speech class as the training datasets are highly imbalanced. In this paper, we propose HateGAN, a deep generative reinforcement learning model, which addresses the challenge of imbalance class by augmenting the dataset with hateful tweets. We conduct extensive experiments to augment two commonly-used hate speech detection datasets with the HateGAN generated tweets. Our …

On The Generation, Structure, And Semantics Of Grammar Patterns In Source Code Identifiers, Christian D. Newman,, Reem S. Alsuhaibani, Michael J. Decker, Anthony Peruma, Dishant Kaushik, Mohamed Wiem Mkaouer, Emily Hill

Articles

Identifier names are the atoms of program comprehension. Weak identifier names decrease developer productivity and degrade the performance of automated approaches that leverage identifier names in source code analysis; threatening many of the advantages which stand to be gained from advances in artificial intelligence and machine learning. Therefore, it is vital to support developers in naming and renaming identifiers. In this paper, we extend our prior work, which studies the primary method through which names evolve: rename refactorings. In our prior work, we contextualize rename changes by examining commit messages and other refactorings. In this extension, we further consider data …

Generating Concept Based Api Element Comparison Using A Knowledge Graph, Yang Liu, Mingwei Liu, Xin Peng, Christoph Treude, Zhenchang Xing, Xiaoxin Zhang

Research Collection School Of Computing and Information Systems

Recommender systems are a valuable tool for software engineers. For example, they can provide developers with a ranked list of files likely to contain a bug, or multiple auto-complete suggestions for a given method stub. However, the way these recommender systems interact with developers is often rudimentary—a long list of recommendations only ranked by the model’s confidence. In this vision paper, we lay out our research agenda for re-imagining how recommender systems for software engineering communicate their insights to developers. When issuing recommendations, our aim is to recommend diverse rather than redundant solutions and present them in ways that highlight …

Deep Multi-Task Learning For Depression Detection And Prediction In Longitudinal Data, Guansong Pang, Ngoc Thien Anh Pham, Emma Baker, Rebecca Bentley, Anton Van Den Hengel

Research Collection School Of Computing and Information Systems

Depression is among the most prevalent mental disorders, affecting millions of people of all ages globally. Machine learning techniques have shown effective in enabling automated detection and prediction of depression for early intervention and treatment. However, they are challenged by the relative scarcity of instances of depression in the data. In this work we introduce a novel deep multi-task recurrent neural network to tackle this challenge, in which depression classification is jointly optimized with two auxiliary tasks, namely one-class metric learning and anomaly ranking. The auxiliary tasks introduce an inductive bias that improves the classification model's generalizability on small depression …

Is2020 A Competency Model For Undergraduate Programs In Information Systems: The Joint Acm/Ais Is2020 Task Force, Paul Leidig, Hannu Salmela

College of Computing Peer-Reviewed Articles

The IS2020 report is the latest in a series of model curricula recommendations and guidelines for undergraduate degrees in Information Systems (IS). The report builds on the foundations developed in previous model curricula reports to develop a major revision of the model curriculum with the inclusion of significant new characteristics. Specifically, the IS2020 report does not directly prescribe a degree structure that targets a specific context or environment. Rather, the IS2020 report provides guidance regarding the core content of the curriculum that should be present but also provides flexibility to customize curricula according to local institutional needs.

Politeness In Security Directives: Insights In Browser Compliance For The Human Element, Deanna House, Gabe Giordano

Information Systems and Quantitative Analysis Faculty Publications

The technical protection provided by information security technology is necessary as a frontline defense against threats. However, the human element adds great risk to systems and cannot be ignored. This research explores the human elements related to security communications and intention to comply with security directives. Security communications are more-commonly being sent using richer computer-based channels. While the goal of security communications is the gain compliance, there is still much to learn related to what influences a user to comply. This research explores the effects that (im)politeness has on intention to comply with security directives. The research utilized an experiment …

Exploring The Potential Of Defeasible Argumentation For Quantitative Inferences In Real-World Contexts: An Assessment Of Computational Trust, Lucas Rizzo, Pierpaolo Dondio, Luca Longo

Articles

Argumentation has recently shown appealing properties for inference under uncertainty and conflicting knowledge. However, there is a lack of studies focused on the examination of its capacity of exploiting real-world knowledge bases for performing quantitative, case-by-case inferences. This study performs an analysis of the inferential capacity of a set of argument-based models, designed by a human reasoner, for the problem of trust assessment. Precisely, these models are exploited using data from Wikipedia, and are aimed at inferring the trustworthiness of its editors. A comparison against non-deductive approaches revealed that these models were superior according to values inferred to recognised trustworthy …

Debunking Rumors On Twitter With Tree Transformer, Jing Ma, Wei Gao

Research Collection School Of Computing and Information Systems

Rumors are manufactured with no respect for accuracy, but can circulate quickly and widely by "word-of-post" through social media conversations. Conversation tree encodes important information indicative of the credibility of rumor. Existing conversation-based techniques for rumor detection either just strictly follow tree edges or treat all the posts fully-connected during feature learning. In this paper, we propose a novel detection model based on tree transformer to better utilize user interactions in the dialogue where post-level self-attention plays the key role for aggregating the intra-/inter-subtree stances. Experimental results on the TWITTER and PHEME datasets show that the proposed approach consistently improves …

A Hybrid Approach For Detecting Prerequisite Relations In Multi-Modal Food Recipes, Liangming Pan, Jingjing Chen, Shaoteng Liu, Chong-Wah Ngo, Min-Yen Kan, Tat-Seng Chua

Research Collection School Of Computing and Information Systems

Modeling the structure of culinary recipes is the core of recipe representation learning. Current approaches mostly focus on extracting the workflow graph from recipes based on text descriptions. Process images, which constitute an important part of cooking recipes, has rarely been investigated in recipe structure modeling. We study this recipe structure problem from a multi-modal learning perspective, by proposing a prerequisite tree to represent recipes with cooking images at a step-level granularity. We propose a simple-yet-effective two-stage framework to automatically construct the prerequisite tree for a recipe by (1) utilizing a trained classifier to detect pairwise prerequisite relations that fuses …

Jointly Optimizing Sensing Pipelines For Multimodal Mixed Reality Interaction, Ramesh Darshana Rathnayake Kanatta Gamage, Ashen De Silva, Dasun Puwakdandawa, Lakmal Meegahapola, Archan Misra, Indika Perera

Research Collection School Of Computing and Information Systems

Natural human interactions for Mixed Reality Applications are overwhelmingly multimodal: humans communicate intent and instructions via a combination of visual, aural and gestural cues. However, supporting low-latency and accurate comprehension of such multimodal instructions (MMI), on resource-constrained wearable devices, remains an open challenge, especially as the state-of-the-art comprehension techniques for each individual modality increasingly utilize complex Deep Neural Network models. We demonstrate the possibility of overcoming the core limitation of latency–vs.–accuracy tradeoff by exploiting cross-modal dependencies–i.e., by compensating for the inferior performance of one model with an increased accuracy of more complex model of a different modality. We present a …

Attentional Parsing Networks, Marcus Karr

Master's Theses

Convolutional neural networks (CNNs) have dominated the computer vision field since the early 2010s, when deep learning largely replaced previous approaches like hand-crafted feature engineering and hierarchical image parsing. Meanwhile transformer architectures have attained preeminence in natural language processing, and have even begun to supplant CNNs as the state of the art for some computer vision tasks.

This study proposes a novel transformer-based architecture, the attentional parsing network, that reconciles the deep learning and hierarchical image parsing approaches to computer vision. We recast unsupervised image representation as a sequence-to-sequence translation problem where image patches are mapped to successive layers …

Language-Driven Region Pointer Advancement For Controllable Image Captioning, Annika Lindh, Robert J. Ross, John D. Kelleher

Conference papers

Controllable Image Captioning is a recent sub-field in the multi-modal task of Image Captioning wherein constraints are placed on which regions in an image should be described in the generated natural language caption. This puts a stronger focus on producing more detailed descriptions, and opens the door for more end-user control over results. A vital component of the Controllable Image Captioning architecture is the mechanism that decides the timing of attending to each region through the advancement of a region pointer. In this paper, we propose a novel method for predicting the timing of region pointer advancement by treating the …

Jointly Optimizing Sensing Pipelines For Multimodal Mixed Reality Interaction, Darshana Rathnayake, Ashen De Silva, Dasun Puwakdandawa, Lakmal Meegahapola, Archan Misra, Indika Perera

Research Collection School Of Computing and Information Systems

Natural human interactions for Mixed Reality Applications are overwhelmingly multimodal: humans communicate intent and instructions via a combination of visual, aural and gestural cues. However, supporting low-latency and accurate comprehension of such multimodal instructions (MMI), on resource-constrained wearable devices, remains an open challenge, especially as the state-of-the-art comprehension techniques for each individual modality increasingly utilize complex Deep Neural Network models. We demonstrate the possibility of overcoming the core limitation of latency--vs.--accuracy tradeoff by exploiting cross-modal dependencies -- i.e., by compensating for the inferior performance of one model with an increased accuracy of more complex model of a different modality. We …

Adelaidecyc At Semeval-2020 Task 12: Ensemble Of Classifiers For Offensive Language Detection In Social Media, Mahen Herath, Thushari Atapattu, Hoang Anh Dung, Christoph Treude, Katrina Falkner

Research Collection School Of Computing and Information Systems

This paper describes the systems our team (AdelaideCyC) has developed for SemEval Task 12 (OffensEval 2020) to detect offensive language in social media. The challenge focuses on three subtasks – offensive language identification (subtask A), offense type identification (subtask B), and offense target identification (subtask C). Our team has participated in all the three subtasks. We have developed machine learning and deep learning-based ensembles of models. We have achieved F1-scores of 0.906, 0.552, and 0.623 in subtask A, B, and C respectively. While our performance scores are promising for subtask A, the results demonstrate that subtask B and C still …

Argumentation Stance Polarity And Intensity Prediction And Its Application For Argumentation Polarization Modeling And Diverse Social Connection Recommendation, Joseph Winstead Sirrianni

Graduate Theses and Dissertations

Cyber argumentation platforms implement theoretical argumentation structures that promote higher quality argumentation and allow for informative analysis of the discussions. Dr. Liu’s research group has designed and implemented a unique platform called the Intelligent Cyber Argumentation System (ICAS). ICAS structures its discussions into a weighted cyber argumentation graph, which describes the relationships between the different users, their posts in a discussion, the discussion topic, and the various subtopics in a discussion. This platform is unique as it encodes online discussions into weighted cyber argumentation graphs based on the user’s stances toward one another’s arguments and ideas. The resulting weighted cyber …

Improving Gan Training With Probability Ratio Clipping And Sample Reweighting, Yue Wu, Pan Zhou, Andrew Wilson Gordon, Eric Xing, Zhiting Hu

Research Collection School Of Computing and Information Systems

Despite success on a wide range of problems related to vision, generative adversarial networks (GANs) often suffer from inferior performance due to unstable training, especially for text generation. To solve this issue, we propose a new variational GAN training framework which enjoys superior training stability. Our approach is inspired by a connection of GANs and reinforcement learning under a variational perspective. The connection leads to (1) probability ratio clipping that regularizes generator training to prevent excessively large updates, and (2) a sample re-weighting mechanism that improves discriminator training by downplaying bad-quality fake samples. Moreover, our variational GAN framework can provably …

Interventional Few-Shot Learning, Zhongqi Yue, Zhang Hanwang, Qianru Sun, Xian-Sheng Hua

Research Collection School Of Computing and Information Systems

We uncover an ever-overlooked deficiency in the prevailing Few-Shot Learning (FSL) methods: the pre-trained knowledge is indeed a confounder that limits the performance. This finding is rooted from our causal assumption: a Structural Causal Model (SCM) for the causalities among the pre-trained knowledge, sample features, and labels. Thanks to it, we propose a novel FSL paradigm: Interventional Few-Shot Learning (IFSL). Specifically, we develop three effective IFSL algorithmic implementations based on the backdoor adjustment, which is essentially a causal intervention towards the SCM of many-shot learning: the upper-bound of FSL in a causal view. It is worth noting that the contribution …

A Bert-Based Dual Embedding Model For Chinese Idiom Prediction, Minghuan Tan, Jing Jiang

Research Collection School Of Computing and Information Systems

Chinese idioms are special fixed phrases usually derived from ancient stories, whose meanings are oftentimes highly idiomatic and non-compositional. The Chinese idiom prediction task is to select the correct idiom from a set of candidate idioms given a context with a blank. We propose a BERT-based dual embedding model to encode the contextual words as well as to learn dual embeddings of the idioms. Specifically, we first match the embedding of each candidate idiom with the hidden representation corresponding to the blank in the context. We then match the embedding of each candidate idiom with the hidden representations of all …

Semantic And Network Evolution In Crowdsourced Idea Refinement, Zhi Wei Vincent Mack, Qian Tang

Research Collection School Of Computing and Information Systems

Successful ideation is a cornerstone of development. Advancements in information technology have allowed firms to harness collective intelligence through crowdsourcing ideation. New ideas often undergo iterative development and evaluation during a refinement phase, and only high-quality ideas get accepted for implementation. Therefore, our paper studies how idea refinement influences idea quality from both the content semantics and network structure perspectives, in the context of the online ideation community of TVTropes. For the content revisions of the idea, we examine the influence of semantic stability and convergence. For the network evolution of the idea, we investigate the impacts of refined network …

Computational Cognition And Deep Learning, Andy Malinsky

The Compass

No abstract provided.

Covid-19 And Mental Health/Substance Use Disorders On Reddit: A Longitudinal Study, Amanuel Alambo, Swati Padhee, Tanvi Banerjee, Krishnaprasad Thirunarayan

Computer Science and Engineering Faculty Publications

COVID-19 pandemic has adversely and disproportionately impacted people suffering from mental health issues and substance use problems. This has been exacerbated by social isolation during the pandemic and the social stigma associated with mental health and substance use disorders, making people reluctant to share their struggles and seek help. Due to the anonymity and privacy they provide, social media emerged as a convenient medium for people to share their experiences about their day to day struggles. Reddit is a well-recognized social media platform that provides focused and structured forums called subreddits, that users subscribe to and discuss their experiences with …

Energy-Based Neural Modelling For Large-Scale Multiple Domain Dialogue State Tracking, Anh Duong Trinh, Robert J. Ross, John D. Kelleher

Conference papers

Scaling up dialogue state tracking to multiple domains is challenging due to the growth in the number of variables being tracked. Furthermore, dialog state tracking models do not yet explicitly make use of relationships between dialogue variables, such as slots across domains. We propose using energy-based structure prediction methods for large-scale dialogue state tracking task in two multiple domain dialogue datasets. Our results indicate that: (i) modelling variable dependencies yields better results; and (ii) the structured prediction output aligns with the dialogue slot-value constraint principles. This leads to promising directions to improve state-of-the-art models by incorporating variable dependencies into their …

Bert Efficacy On Scientific And Medical Datasets: A Systematic Literature Review, Clayton Cohn

College of Computing and Digital Media Dissertations

Bidirectional Encoder Representations from Transformers (BERT) [Devlin et al., 2018] has been shown to be effective at modeling a multitude of datasets across a wide variety of Natural Language Processing (NLP) tasks; however, little research has been done regarding BERT’s effectiveness at modeling domain-specific datasets. Specifically, scientific and medical datasets present a particularly difficult challenge in NLP, as these types of corpora are often rife with technical jargon that is largely absent from the canonical corpora that BERT and other transfer learning models were originally trained on. This thesis is a Systematic Literature Review (SLR) of twenty-seven studies that were …

An Analysis Of The Effectiveness Of Cuban Cyberactivism, Neta Kanny

Journal of Global Initiatives: Policy, Pedagogy, Perspective

This article applies power-law degree distribution and network theory as a conceptual framework for assessing the effectiveness of Cuban cyberactivism based on an analysis of previous scholarly work on the topic. While Cuban cyberactivism indicates the potential for a more accessible, just, and transparent media environment, the movement continues to face serious obstacles due to extensive controls put in place by the Communist Party of Cuba.

Deep Learning In The Maintenance Industry, Paulo Cesar Ribeiro Silva

Doctoral

No abstract provided.

Countermeasures Against Various Network Attacks Using Machine Learning Methods, Yi Li

USF Tampa Graduate Theses and Dissertations

With the rapid development of a computer network, our lives are already inseparable from it. Wireless Fidelity (Wi-Fi) is in use everywhere; more and more devices are connected to the Internet, and many companies and individuals tend to store their data and information online. Furthermore, it is now very convenient to communicate with each other through email and text messages. However, widespread networks also provide more attack surfaces for attackers. There are a variety of network attacks aimed at information theft. To better defend against those network attacks, one needs to have a broad knowledge of existing attacks. In this …

Topics And Challenges In Teaching Computer Science For Gifted And Talented Students, Tom Meyer, Namrata Pandya

Faculty Publications & Research

No abstract provided.

Literature Retrieval For Precision Medicine With Neural Matching And Faceted Summarization, Jiho Noh, Ramakanth Kavuluru

Institute for Biomedical Informatics Faculty Publications

Information retrieval (IR) for precision medicine (PM) often involves looking for multiple pieces of evidence that characterize a patient case. This typically includes at least the name of a condition and a genetic variation that applies to the patient. Other factors such as demographic attributes, comorbidities, and social determinants may also be pertinent. As such, the retrieval problem is often formulated as ad hoc search but with multiple facets (e.g., disease, mutation) that may need to be incorporated. In this paper, we present a document reranking approach that combines neural query-document matching and text summarization toward such retrieval scenarios. Our …

Tangi: Tangible Proxies For Embodied Object Exploration And Manipulation In Virtual Reality, Martin Feick, Scott Bateman, Anthony Tang, Anthony Tang

Research Collection School Of Computing and Information Systems

Exploring and manipulating complex virtual objects is challenging due to limitations of conventional controllers and free-hand interaction techniques. We present the TanGi toolkit which enables novices to rapidly build physical proxy objects using Composable Shape Primitives. TanGi also provides Manipulators allowing users to build objects including movable parts, making them suitable for rich object exploration and manipulation in VR. With a set of different use cases and applications we show the capabilities of the TanGi toolkit and evaluate its use. In a study with 16 participants, we demonstrate that novices can quickly build physical proxy objects using the Composable Shape …

Coupled Hierarchical Transformer For Stance-Aware Rumor Verification In Social Media Conversations, Jianfei Yu, Jing Jiang, Ling Min Serena Khoo, Hai Leong Chieu, Rui Xia

Research Collection School Of Computing and Information Systems

The prevalent use of social media enables rapid spread of rumors on a massive scale, which leads to the emerging need of automatic rumor verification (RV). A number of previous studies focus on leveraging stance classification to enhance RV with multi-task learning (MTL) methods. However, most of these methods failed to employ pre-trained contextualized embeddings such as BERT, and did not exploit inter-task dependencies by using predicted stance labels to improve the RV task. Therefore, in this paper, to extend BERT to obtain thread representations, we first propose a Hierarchical Transformer1 , which divides each long thread into shorter subthreads, …