This paper investigates the application of co-training and self-training to word sense disambiguation. Optimal and empirical parameter selection methods for co-training and self-training are investigated, with various degrees of error reduction. A new method that combines co-training with majority voting is introduced, with the effect of smoothing the bootstrapping learning curves, and improving the average performance.

This paper describes a co-training approach that uses the text and citation information of a research article as two different views to predict the topic of an article.

Paper on co-training over domain-independent and domain-dependent features for sentiment analysis of an online cancer support community.

Data management plan for the grant "Collaborative Research: CCRI: Planning: A Multilayer Network (MLN) Community Infrastructure for Data,Interaction,Visualization, and softwarE(MLN-DIVE)." Research relating to creating a community infrastructure for researchers using multilayer networks (MLN). This project uses a formally established network decoupling approach to perform various aggregate analysis (community, centrality, substructure detection, etc.) using individual layers and composing them. The broader impact of this planning project is to provide meaningful and appropriate analysis tools that are grounded in theory to a broad range of applications from different domains. The focus is on facilitating the mainstream use of multilayer network analysis in data analysis, research and teaching.

Data management plan for the grant, "Collaborative Research: CyberTraining: Pilot: Research Workforce Development for Deep Learning Systems in Advanced GPU Cyberinfrastructure." This project aims to develop a novel set of interactive training materials, including hands-on lecture modules, invited research talks from renowned researchers, and an interdisciplinary collaborative project in an intensive workshop, integrating a wide variety of advanced and inter-connected techniques employed by research workforce for deep learning (DL) systems in advanced GPU cyberinfrastructure (CI). Specifically, this project focuses on training seniors, graduate students, and researchers on how advanced GPU CI can be efficiently utilized and improved to enable high-performance DL systems for data-intensive DL applications in geoscience (GS) and computer science and engineering (CSE) research. The goal is to foster future CI users and contributors to adopt, develop, and improve advanced GPU CI for DL systems in their research.

Data management plan for the grant, "Collaborative Research: Engaging Blind and Visually Impaired Youth in Computer Science through Music Programming."

Data management for the grant, "Collaborative Research: Framework Implementations: CSSI: CANDY: Cyberinfrastructure for Accelerating Innovation in Network Dynamics." Research addressing the lack of a comprehensive cyberinfrastructure that supports innovative research challenges in large-scale, complex, dynamic networks by developing a novel platform, called CANDY (Cyberinfrastructure for Accelerating Innovation in Network Dynamics), based on efficient, scalable parallel algorithm design for dynamic networks and high-performance software development with performance optimization.

Data management plan for the grant, "Collaborative Research: IRES Track I: Wireless Federated Fog Computing for Remote Industry 4.0 Applications."

Data management plan for the grant, "Collaborative Research: SaTC: CORE: Small: Privacy protection of Vehicles location in Spatial Crowdsourcing under realistic adversarial models." Research to develop new location privacy protection techniques by considering vehicles’ mobility features in the road network, and consequently lead to a more secure and trustworthy computing environment in location-based services (LBSs). As privacy concerns are still among the main obstacles for mobile users to participate in many advanced LBSs, this project is poised to contribute to the wider adoption of LBSs for many applications (e.g. navigation systems and location-based recommendation systems). The project will also provide a set of diverse and interesting topics for undergraduate and graduate students and outreach activities for the community.

Data management plan for the grant, "Collaborative Research: SHF: Small: LEGAS: Learning Evolving Graphs At Scale." This project aims to improve the computation efficiency of graph neural networks (GNNs), which are an emerging class of deep learning models on graphs, with many successful applications, such as, recommendation systems, drug discovery, social network analysis, and code vulnerability detection. This project aims to to design an efficient GNN framework via algorithm and system co-design for both static and dynamic graphs.

Article on combining hashing and abstraction in sparse high dimensional feature spaces.

This paper discusses combining lexical resources for contextual synonym expansion.

This article discusses the effectiveness of utilizing local spatial variations in environmental data to uncover the statistical relationships between West Nile Virus (WNV) risk and environmental factors.

Poster for the 2014 MidSouth Computational Biology and Bioinformatics Society (MCBIOS) Annual Conference. This poster discusses a comparison of machine learning algorithms for identifying cancer types.

This paper discusses automatic recognition of humor.

In this paper, the authors explore several computational models for incongruity resolution.

This article is a study where the authors integrated the multi-omics datasets such as copy number aberration, DNA methylation, gene and microRNA expression to identify the signature microRNA-gene associations from frequently aberrated DNA regions across pan-cancer utilizing a LASSO-based regression approach.

Video from the Fall 2018 3 Minute Thesis (3MT®) Final Competition. In this video, Shraddha Piparia presents her research methods, findings, and its significance in non-technical language.

This article discusses corpus-based and knowledge-based measures of text semantic similarity.

This paper discusses how to locate emotions.

This paper describes on-going efforts to annotate a corpus of almost 16,000 answer pairs with an estimated 69,000 fine-grained entailment relationships.

Article introduces a large corpus of metaphor novelty scores for syntactically related word pairs, and releases it freely to the research community. This article describes the corpus, includes an analysis of its score distribution and the types of word pairs included in the corpus, and provides a brief overview of standard metaphor detection corpora.

Article presenting a corpus of negations and their underlying positive interpretations using negations from Simple Wikipedia, automatically generating potential positive interpretations, and collecting manual annotations that effectively rewrite the negation in positive terms. This article was presented at the Eighth Joint Conference on Lexical and Computational Semantics (SEM 2019) in Minneapolis, Minnesota, June 6-7, 2019.

Article studying the problem of incorporating corpus, ontology, and semantic predications to learn the embeddings of MeSH terms. The authors propose a novel framework, Corpus, Ontology, and Semantic predications-based MeSH term embedding (COS), to generate high-quality MeSH term embeddings.