Kevin Hyekang Joo

Kevin Hyekang Joo

About Me | I previously obtained an M.S. in Computer Science from the University of Maryland, College Park, advised by Prof. David W. Jacobs. My primary research interest is Multimodal Affective Computing, via minimal human expenses externally.

Activities | Outside of research, I like to travel around, dine outside, and listen to music. In addition, as a supporter of diversity, equity, and inclusion, I like to learn about the needs of the underrepresented minorities and explore the potential solutions to those problems with my skills.

*I am currently looking for a Ph.D. position starting in Fall 2024!*

Previously, I have been fortunate enough to have opportunities to conduct research on Temporal Action Proposal Generation and Anomalous Activity Detection with Prof. Ngan Le at the University of Arkansas and on ML-based Scalable Optimization with Prof. Leilani Battle from the Interactive Data Lab at the University of Washington.

directory

email

kevinjoo (at) andrew.cmu.edu (Primary)
   hkjoo    (at) cs.umd.edu (UMD Inquiries)
   hkjoo    (at) cs.washington.edu (UW Inquiries)
   hj025    (at) uark.edu (UArk Inquiries)

resume

github

scholar

News

6/23: Starts Research Internship at MIT Lincoln Lab in Lexington, MA
5/23: Graduates from the University of Maryland with an M.S. in Computer Science, advised by Prof. David W. Jacobs
5/22: Starts Research Internship with Prof. Ngan Le at the University of Arkansas over the summer
5/22: Serves as a part-time teaching assistant for Google Applied Machine Learning Intensive Summer Program
2/22: A paper [3] accepted at ACM SIGMOD 2022
12/21: Serves as a Student Committee Member for Graduate School Admissions for the Department of Computer Science at the University of Maryland
10/21: A paper [2] accepted at BMVC 2021 for an oral presentation
7/21: Selected as one of 12 receipients of $10,000 Google Scholarship in the year of 2021, funded by Google, for my postgraduate studies in Computer Science
6/21: Starts Research Internship with Prof. Leilani Battle at the University of Washington through the NSF-funded Research Experiences for Undergraduates (REU)
5/21: Graduates from the University of Maryland with a B.S. in Computer Science (Machine Learning Track) and a Minor in Technology Entrepreneurship
6/20: Starts Software Engineering Internship at Amazon Devices (Remote)
6/19: Starts Software Engineering Internship at Amazon Web Services (AWS) in Herndon, VA

Publications

	[4] CLIP-TSA: CLIP-Assisted Temporal Self-Attention for Weakly-Supervised Video Anomaly Detection new Hyekang Kevin Joo, Khoa Vo, Kashu Yamazaki, Ngan Le Under Review computer vision tags: video anomaly detection, activity analysis, vision-language, clip, temporal self-attention pdf / bibtex In this paper, the magnitude-based video anomaly detection model is optimized in a weakly-supervised manner. The proposed training pipeline notably leverages the ViT-encoded visual features from CLIP, in contrast with the conventional C3D or I3D features in the domain, to extract discriminative representations efficiently and nominates the snippets of interest with the proposed Temporal Self-Attention.
	[3] Demonstration of VegaPlus: Optimizing Declarative Visualization Languages Junran Yang, Hyekang Kevin Joo, Sai S. Yerramreddy, Siyao Li, Dominik Moritz, Leilani Battle ACM Special Interest Group on Management of Data (SIGMOD) Conference, 2022 machine learning hci ACM SIGMOD Demo '22 statistics [h5-index: 68]: 33.8% accept tags: scalable optimization, data visualization, dataflow, machine learning conference / video / github / pdf / bibtex We propose a system that takes declarative specifications to automatically offload computation-intensive processing to a separate DBMS by optimizing visualization plans through machine learning techniques. In this project, I was primarily in charge of scalable optimization aspect of the project. [ Note: This is the UW REU project* that I worked on with Profs. Battle and Moritz in Summer 2021. Keep an eye out for the extended version of the work, to be done and submitted in Fall 2023!* ]
	[2] AEI: Actors-Environment Interaction with Adaptive Attention for Temporal Action Proposals Generation Khoa Vo, Hyekang Joo, Kashu Yamazaki, Sang Truong, Kris Kitani, Minh-Triet Tran, Ngan Le The British Machine Vision Conference (BMVC), 2021 (Oral Presentation) computer vision BMVC '21 statistics [h5-index: 75]: 40 (3.33%) oral* / 437 (36.23%) accept / 1206 submissions* tags: temporal action proposal generation, activitynet conference / github / pdf / bibtex The human perceives the establishment of an action in a video through the interaction between an actor and the surrounding environment. We borrow the idea in producing the proposed Actor-Environment Interaction network.

	[1] Guided Hyperparameter Tuning Through Visualization and Inference Hyekang Joo, Calvin Bao, Ishan Sen, Furong Huang, Leilani Battle arXiv, 2021 machine learning hci tags: hyperparameter tuning, AutoML pdf / bibtex We create an AutoML tool and run evaluations on machine translation and image classification through user studies.

*PI & Co-PI are marked green on mouse-hover

**My legal name on the U.S. Passport is Kevin Hyekang Joo, but I occasionally put my name as Hyekang Kevin Joo on papers.

Service

UMDCS

Graduate School Admissions

Department of Computer Science
University of Maryland, College Park

Student Committee Member for CS Graduate School Admissions
Spring 2022

UMDACM

Student Groups

Association for Computing Machinery
University of Maryland, College Park

President
2020-2021 2019-2020

Vice President
2021-2022

Secretary
2018-2019

ALD

Alpha Lambda Delta Honor Society
University of Maryland, College Park

Webmaster
*2022-2023 **2021-2022 2020-2021 2019-2020

*Serves as an Auxiliary Co-Webmaster
**Received the prestigious Order of the Torch Award of the Year '22

cybersecurity

Cybersecurity Club
University of Maryland, College Park

Secretary
2018-2019

google-ed

Mentorship

Google Applied Machine Learning Intensive Summer Program
University of Arkansas, Fayetteville

Teaching Assistant
Summer 2022
After partnering with the National Action Council for Minorities in Engineering and Google, the College of Engineering at the University of Arkansas is hosting the Google Applied Machine Learning Intensive Bootcamp for 10 weeks over the summer to introduce undergraduate (URM-focused) college students to applied machine learning concepts.

i4c

Diversity & Inclusion in Computing Education (DICE)
University of Maryland, College Park

Mentor
2022-2023 2021-2022

bitcamp

IEEE/CVF CVPR 2022
New Orleans Ernest N. Morial Convention Center

Mentor for the CVPR Academy (HBCU/MI Workshop)
Summer 2022

diversity

Diversity, Equity & Inclusion (DE&I) Conference

Tapia Conference
2022 2021
Grace Hopper Conference
2022 2021
CSUN Assistive Technology Conference
2022
CRA Grad Cohort Workshop for IDEALS:
Inclusion, Diversity, Equity, Accessibility, and Leadership Skills
2023 2022

Teaching Assistant Experience

UMDCS

Computer Science TA @ UMD

Computer Vision (CMSC426)
Instructor: Dr. Mohammad Nayeem Teli
Spring 2022
Natural Language Processing (MSML641)
Instructor: Dr. Shabnam Tafreshi
Spring 2022
Object-Oriented Programming I (CMSC131)
Instructor: Senior Lecturer Nelson Padua-Perez
*Fall 2022 *Spring 2020 *Fall 2019
*Led two discussion sessions in each semester as a Discussion Teacher TA

Why so enthusiastic about being a Teacher TA?

Knowledge is meant to be shared! I enjoy interacting with people and sharing my knowledge with others! Oftentimes, I serve as a guest speaker, too, at academic clubs, dicussing my research, industry, and academic backgrounds with fellow students! :)

Here are seven other reasons I strongly prefer to be a Teaching TA for CMSC131:

1. Greatly improves my skills for oral presentation
2. Forms social connections with newcomers
3. Excellent memory refresher of Java
4. Enables self-evaluation of my communication skill
5. Student-evaluation of my skills at the end of sem
6. Builds up confidence as a speaker
7. Solid setting for practicing my leadership,
organization, and on-the-fly Q&A skills

hinman

Technology & Engineering Entrepreneurship TA @ UMD

Entrepreneurial Opportunity Analysis and Decision-Making in Technology Ventures (ENES210)
Instructor: Prof. James Green
Spring 2021
Marketing High-Technology Products and Innovations (ENES462)
Instructor: Lecturer Michael Pratt
Spring 2021
Leading and Financing the Technology Venture (ENES466)
Instructor: Lecturer Michael Pratt
Spring 2021

Undergraduate Internship Experience

amazon

Amazon

Amazon Devices
Denver, CO (Remote Internship)
Summer 2020

Amazon Web Services (AWS)
Herndon, VA
Summer 2019

0xC0C4C0DA
(It's not a typical soda!)

HKJ

v23.5.21

My Erdös number is 4
[ Erdös - P. O'Neil - M. Stonebraker - L. Battle - Me ]

More About Me :) Special Thanks