Computer Information and Graphics Technology Works

Permanent URI for this collection

Browse

Recent Submissions

Now showing 1 - 10 of 76
  • Item
    Methods of Current Knowledge Teaching on the Cybersecurity Example
    (MDPI, 2022-10-22) Nyemkova, Elena; Justice, Connie; Liaskovska, Solomiia; Lakh, Yuriy; Computer Information and Graphics Technology, Purdue School of Engineering and Technology
    Teaching of modern cybersecurity specialists should be up to date and use the newest methods and methodologies in universities as the IT industry is rapidly growing and constantly changing. A good idea is to use methods of management in IT companies as methods for current knowledge teaching of university students. It is also worth engaging students not only in educational international projects but the research projects as well. This work analyzes the method for teaching students, and the Scrum methodology was selected and implemented for educational and research projects. Students participated in both projects, however, Scrum models should be different for them and this is illustrated in the paper. The visualization of collected statistical data of the performed educational project illustrated distributions of students by specialization and by marks. The distributions by marks showed that using the Scrum model for the teaching course significantly increases the marks compared with the average level marks of the students in their specializations.
  • Item
    Maximum Density Divergence for Domain Adaptation
    (IEEE, 2021) Li, Jingjing; Chen, Erpeng; Ding, Zhengming; Zhu, Lei; Lu, Ke; Shen, Heng Tao; Computer Information and Graphics Technology, Purdue School of Engineering and Technology
    Unsupervised domain adaptation addresses the problem of transferring knowledge from a well-labeled source domain to an unlabeled target domain where the two domains have distinctive data distributions. Thus, the essence of domain adaptation is to mitigate the distribution divergence between the two domains. The state-of-the-art methods practice this very idea by either conducting adversarial training or minimizing a metric which defines the distribution gaps. In this paper, we propose a new domain adaptation method named adversarial tight match (ATM) which enjoys the benefits of both adversarial training and metric learning. Specifically, at first, we propose a novel distance loss, named maximum density divergence (MDD), to quantify the distribution divergence. MDD minimizes the inter-domain divergence ("match" in ATM) and maximizes the intra-class density ("tight" in ATM). Then, to address the equilibrium challenge issue in adversarial domain adaptation, we consider leveraging the proposed MDD into adversarial domain adaptation framework. At last, we tailor the proposed MDD as a practical learning loss and report our ATM. Both empirical evaluation and theoretical analysis are reported to verify the effectiveness of the proposed method. The experimental results on four benchmarks, both classical and large-scale, show that our method is able to achieve new state-of-the-art performance on most evaluations.
  • Item
    Weakly-Supervised Cross-Domain Adaptation for Endoscopic Lesions Segmentation
    (IEEE, 2021) Dong, Jiahua; Cong, Yang; Sun, Gan; Yang, Yunsheng; Xu, Xiaowei; Ding, Zhengming; Computer Information and Graphics Technology, Purdue School of Engineering and Technology
    Weakly-supervised learning has attracted growing research attention on medical lesions segmentation due to significant saving in pixel-level annotation cost. However, 1) most existing methods require effective prior and constraints to explore the intrinsic lesions characterization, which only generates incorrect and rough prediction; 2) they neglect the underlying semantic dependencies among weakly-labeled target enteroscopy diseases and fully-annotated source gastroscope lesions, while forcefully utilizing untransferable dependencies leads to the negative performance. To tackle above issues, we propose a new weakly-supervised lesions transfer framework, which can not only explore transferable domain-invariant knowledge across different datasets, but also prevent the negative transfer of untransferable representations. Specifically, a Wasserstein quantified transferability framework is developed to highlight wide-range transferable contextual dependencies, while neglecting the irrelevant semantic characterizations. Moreover, a novel self-supervised pseudo label generator is designed to equally provide confident pseudo pixel labels for both hard-to-transfer and easy-to-transfer target samples. It inhibits the enormous deviation of false pseudo pixel labels under the self-supervision manner. Afterwards, dynamically-searched feature centroids are aligned to narrow category-wise distribution shift. Comprehensive theoretical analysis and experiments show the superiority of our model on the endoscopic dataset and several public datasets.
  • Item
    A Computationally Effective Pedestrian Detection using Constrained Fusion with Body Parts for Autonomous Driving
    (IEEE, 2021) Islam, Muhammad Mobaidul; Newaz, Abdullah Al Redwan; Tian, Renran; Homaifar, Abdollah; Karimoddini, Ali; Computer Information and Graphics Technology, School of Engineering and Technology
    This paper addresses the problem of detecting pedestrians using an enhanced object detection method. In particular, the paper considers the occluded pedestrian detection problem in autonomous driving scenarios where the balance of performance between accuracy and speed is crucial. Existing works focus on learning representations of unique persons independent of body parts semantics. To achieve a real-time performance along with robust detection, we introduce a body parts based pedestrian detection architecture where body parts are fused through a computationally effective constraint optimization technique. We demonstrate that our method significantly improves detection accuracy while adding negligible runtime overhead. We evaluate our method using a real-world dataset. Experimental results show that the proposed method outperforms existing pedestrian detection methods.
  • Item
    Incorporating Global Learning Perspectives in a Freshman Computing Curriculum
    (IEEE, 2022-10-11) Mithun, Shamima; Goldfarb, Nancy; Computer Information and Graphics Technology, School of Engineering and Technology
    This research-to-practice full paper describes our integration of global learning perspectives through a research-based group project in a First-Year Seminar course for new technology major students at our urban Midwestern university, IUPUI (Indiana University - Purdue University Indianapolis). Since 2003, the ACE (American Council on Education) has emphasized global competencies, which are defined as "the attitudes, skills, and knowledge to live and work in a multicultural and interconnected world". Despite the increasing recognition of the importance of developing these global competencies, opportunities for young people to do so suffer from issues of accessibility. Educational institutions are challenged with providing opportunities to prepare students for global citizenship in the twenty-first century and are working to expand global competency education. Our undergraduate institution is no exception.In accordance with this mission, we incorporated global learning perspectives through a group project in our First-Year-Seminar course to increase students’ interest in global learning experiences (such as studying abroad) and provide resources for students to develop global competencies. This is important both for personal development in the quest for a more equitable world and employability; employers repeatedly convey that awareness of global issues is a highly desirable characteristic in potential hires.In our implementation, students selected a global issue, chosen from the United Nations Sustainable Development Goals database, which had unique significance to them and their communities. This approach allowed space for students to take ownership and agency over the content of their learning experiences while ensuring they engaged with the following learning objectives: 1)Team collaboration, communication, and cohesion 2)Conducting independent research on a global problem and its solutions 3)Synthesis of information from multiple sources and perspectives to develop an informed stance 4)Developing a stance regarding a global problem and justification of this stance using data 5)Creating a well-organized deliverable with consideration for the audience (i.e., their peers) and contextWe point to course survey data and student reflections to evaluate our course. Students conveyed how the course structure enabled them to (a) consider global perspectives around issues that may or may not have been salient to them before the course, (b) experience empathy for people experiencing challenges related to the issues of interest and gratitude for their circumstances, and (c) consider their personal role in addressing global issues in their communities. Students also indicated an interest in further addressing such issues through self-education and advocacy on a community and political scale.To further expand efforts to make global competency education accessible, our next implementation will utilize Collaborative Online International Learning experiences in which students virtually collaborate with students outside of the United States through our local Office of International Affairs. Through these collaborations, students will be challenged to consider how such global issues manifest in different communities, cultures, and geographic regions and the implications of these differences for solution design.
  • Item
    Hands-Free Electronic Documentation in Emergency Care Work Through Smart Glasses
    (Springer, 2022-02) Zhang, Zhan; Luo, Xiao; Harris, Richard; George, Susanna; Finkelstein, Jack; Computer Information and Graphics Technology, School of Engineering and Technology
    As U.S. healthcare system moves towards digitization, Electronic Health Records (EHRs) are increasingly adopted by medical providers. However, EHR documentation is not only time-consuming but also difficult to complete in real-time, leading to delayed, missing, or erroneous data entry. This challenge is more evident in time-critical and hands-busy clinical domains, such as Emergency Medical Services (EMS). In recent years, smart glasses have gained momentum in supporting various aspects of clinical care. However, limited research has examined the potential of smart glasses in automating electronic documentation during fast-paced medical work. In this paper, we report the design, development, and preliminary evaluations of a novel system combining smart glasses and EHRs and leveraging natural language processing (NLP) techniques to enable hands-free, real-time documentation in the context of EMS care. Although optimization is needed, our system prototype represents a substantive departure from the status quo in the documentation technology for emergency care providers, and has a high potential to enable real-time documentation while accounting for care providers’ cognitive and physical constraints imposed by the time-critical medical environment.
  • Item
    Novice Versus Advanced Undergraduate Computing Students’ Engagement in Collaboration in an Online Flipped Classroom
    (International Society of the Learning Sciences, 2022) Vickery, Morgan; Mithun, Shamima; Computer Information and Graphics Technology, School of Engineering and Technology
    This paper explores students' engagement in collaborative learning activities within two data science courses (one introductory and one advanced) taught synchronously online during the 2020-2021 academic year. Here, we draw on a multidimensional perspective of student engagement to understand novice and advanced computing students' collaborative learning practices and propose instructional design elements informed by students’ unique needs and the limitations of the virtual format.
  • Item
    Flipped Instructional Design Factors in an Introductory and an Advanced Data Science Course
    (ASEE, 2022-06) Mithun, Shamima; Vickery, Morgan; Luo, Xiao; Computer Information and Graphics Technology, School of Engineering and Technology
    In this full research paper, we evaluate the flipped instructional designs of two undergraduate data science courses at a Midwestern university: an introductory course on database fundamentals and an advanced database design course. This study is built upon our prior work in which we identified a set of eight instructional design factors for effective flipped classrooms in the literature and assessed their efficacy with senior students. Our analysis relies on students’ course evaluations, self-reported survey data, focus group responses, course performance data, and instructor observation data to answer the following research questions: 1. How do the eight instructional design factors for effective flipped classrooms serve novice versus advanced data science students? 2. How should instruction in flipped classrooms be varied for novice versus advanced data science students? Our analysis indicates that novice data science students have different instructional needs and challenges compared to their senior peers, particularly in relation to activities that require peer collaboration and were unmoderated by the instructor. We share the results of our quantitative analysis of self-reported survey data in which students ranked the aforementioned instructional design factors based on their effectiveness for their learning and qualitative analysis which takes student comments (from a free-response survey and focus group data) and instructor observation data to contextualize these rankings and inform our instructional design recommendations. These recommendations address students differing academic and interactional needs within the classroom and are to be implemented within the introductory course in its next iteration: (a) group norming and standardization around expectations for communication/collaboration, (b) transparent disclosure of the learning objectives for each activity, (c) offering guidelines to support students in providing actionable peer feedback, and (d) introducing low-stakes peer evaluations. We conclude with a discussion on the general affordances of the flipped classroom model for both introductory and advanced data science instruction compared to traditional lecture-based approaches.
  • Item
    Electronic Co-design (ECO-design) Workshop for Increasing Clinician Participation in the Design of Health Services Interventions: Participatory Design Approach
    (JMIR Publications, 2022-09-22) Savoy, April; Patel, Himalaya; Shahid, Umber; Offner, Alexis D.; Singh, Hardeep; Giardina, Traber D.; Meyer, Ashley N.D.
    Background: Participation from clinician stakeholders can improve the design and implementation of health care interventions. Participatory design methods, especially co-design methods, comprise stakeholder-led design activities that are time-consuming. Competing work demands and increasing workloads make clinicians' commitments to typical participatory methods even harder. The COVID-19 pandemic further exacerbated barriers to clinician participation in such interventions. Objective: The aim of this study was to explore a web-based participatory design approach to conduct economical, electronic co-design (ECO-design) workshops with primary care clinicians. Methods: We adapted traditional in-person co-design workshops to web-based delivery and adapted co-design workshop series to fit within a single 1-hour session. We applied the ECO-design workshop approach to codevelop feedback interventions regarding abnormal test result follow-up in primary care. We conducted ECO-design workshops with primary care clinicians at a medical center in Southern Texas, using videoconferencing software. Each workshop focused on one of three types of feedback interventions: conversation guide, email template, and dashboard prototype. We paired electronic materials and software features to facilitate participant interactions, prototyping, and data collection. The workshop protocol included four main activities: problem identification, solution generation, prototyping, and debriefing. Two facilitators were assigned to each workshop and one researcher resolved technical problems. After the workshops, our research team met to debrief and evaluate workshops. Results: A total of 28 primary care clinicians participated in our ECO-design workshops. We completed 4 parallel workshops, each with 5-10 participants. We conducted traditional analyses and generated a clinician persona (ie, representative description) and user interface prototypes. We also formulated recommendations for future ECO-design workshop recruitment, technology, facilitation, and data collection. Overall, our adapted workshops successfully enabled primary care clinicians to participate without increasing their workload, even during a pandemic. Conclusions: ECO-design workshops are viable, economical alternatives to traditional approaches. This approach fills a need for efficient methods to involve busy clinicians in the design of health care interventions.
  • Item
    Electronic Health Records’ Support for Primary Care Physicians’ Situation Awareness: A Metanarrative Review
    (Sage, 2023) Savoy, April; Patel, Himalaya; Murphy, Daniel R.; Meyer, Ashley N.D.; Herout, Jennifer; Singh, Hardeep
    Objective: Situation awareness (SA) refers to people's perception and understanding of their dynamic environment. In primary care, reduced SA among physicians increases errors in clinical decision-making and, correspondingly, patients' risk of experiencing adverse outcomes. Our objective was to understand the extent to which electronic health records (EHRs) support primary care physicians (PCPs)' SA during clinical decision-making. Method: We conducted a metanarrative review of papers in selected academic databases, including CINAHL and MEDLINE. Eligible studies included original peer-reviewed research published between January 2012 and August 2020 on PCP-EHR interactions. We iteratively queried, screened, and summarized literature focused on EHRs supporting PCPs' clinical decision-making and care management for adults. Then, we mapped findings to an established SA framework to classify external factors (individual, task, and system) affecting PCPs' levels of SA (1-Perception, 2-Comprehension, and 3-Projection) and identified SA barriers. Results: From 1504 articles identified, we included and synthesized 19 studies. Study designs were largely noninterventional. Studies described EHR workflow misalignments, usability issues, and communication challenges. EHR information, including lab results and care plans, was characterized as incomplete, untimely, or irrelevant. Unmet information needs made it difficult for PCPs to obtain even basic SA, Level 1 SA. Prevalent barriers to PCPs developing SA with EHRs were errant mental models, attentional tunneling, and data overload. Conclusion: Based on our review, EHRs do not support the development of higher levels of SA among PCPs. Review findings suggest SA-oriented design processes for health information technology could improve PCPs' SA, satisfaction, and decision-making.