Sungkyu Shaun Park

Sleep & Mental Health

[A full paper] imputing missing wearable sleep data using AI was published in Sleep, Oct 2023.  
[A full paper] revealing the impact of social dimensions on sleep patterns was published in Scientific Reports, Jun 2023 (#TLDR).  
[A framework paper] for tracking daily well-being using smartwatches was published in IEEE Internet Computing, Aug 2020.  
[A sleep clustering paper] was published in JMIR mHealth & uHealth (JMU), Dec 2019.  

[A sleep prediction work] was presented at KDD'19 (poster session), Aug 2019.  
➥ [Full paper] + [Promotional video clip] for KDD'19.  

[A paper about the effects of child maltreatment on sleep] published in a Korean psychiatric journal (생물치료정신의학), Jun 2022.
[An effectiveness verification paper] on the app Sleeps published in a Korean psychiatric journal (생물치료정신의학), Feb 2021.  

[Science Slam D] I gave a 10-min talk about sleep in Daejeon, Oct 2021 ➥ a news article (in Korean).  
[IBS Interview] My work has been introduced in an interview article at IBS, Mar 2021 (in Korean)!  
[Annual Report] The work has been honorably featured in The 2020 Annual Report of KAIST School of Computing, Feb 2021 (see p. 19, in Korean).  
[Newspaper Column] The work has been introduced in The Korea Economic Daily, Nov 2020.  
[Research Highlight] The work has been introduced in KAIST School of Computing, Oct 2020.  

Close Window

Online Fake News

[One extended abstract] titled "Do Differences in National or Political Identity Matter More for Preferences?" was presented at the poster session of IC2S2 2022 in Jul 2022.

[One paper] titled "An Experimental Study to Understand User Experience and Perception Bias Occurred by Fact-checking Messages" was presented at The Web Conference 2021 in Apr 2021.
[Related research] titled "The Presence of Unexpected Biases in Online Fact-checking" has been published in The Harvard Kennedy School (HKS) Misinformation Review in Jan 2021.

Close Window

Natural Language Processing (NLP)

Another topic modeling paper "Platform-Invariant Topic Modeling via Contrastive Learning to Mitigate Platform-Induced Bias" was presented at The Conference on Empirical Methods in Natural Language Processing (EMNLP 2024) as Findings, Nov 2024.  

A topic modeling paper "Unified Neural Topic Model via Contrastive Learning and Term Weighting" was presented at The 17th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2023), May 2023.  

A topic modeling paper "Contrastive Neural Topic Model Using Term Weighting" won the Outstanding Paper Award at the Conference of Korean Artificial Intelligence Association, 1 August 2022.  

[One paper] titled "COVID-19 Discourse on Twitter: Case Study of Risk Communication in Four Asian Countries" has been accepted at JMIR (Impact Factor = 5.03 [SCIE]), Mar 2021.  
➥ This paper has been featured on PreventionWeb, managed by the UN Office for Disaster Risk Reduction (UNDRR) in Sep 2021! - please see [here].

[A preprint] to explore COVID-19 tweet topics in 4 countries has been released at arXiv, Jun 2020.  
➥ [The same version] has been displayed in Europe PMC, March 2021.  
➥ [A related post] on Medium.  

[An in-depth report] on fake news and infodemic has been released in Donga Science, Dec 2020.  

[A short paper] on a Risk Communication Event Detection Model via contrastive learning has been presented at the 3rd Workshop on NLP for Internet Freedom in COLING 2020, Dec 2020.  

Close Window

Unsupervised Learning

[One paper] titled "Improving Unsupervised Image Clustering With Robust Learning" was presented at CVPR'21 in Jun 2021.

[One paper] titled "Mitigating Embedding and Class Assignment Mismatch in Unsupervised Image Classification" was presented at ECCV'20, Aug 2020.

Close Window

Culture & Fashion

[One paper] titled "Using Web Data to Reveal 22-Year History of Sneaker Designs" was presented at ACM The Web Conference (WWW) 2022, Apr 2022.
➥ This paper was one of the BEST PAPER CANDIDATES (top 3.41% of the accepted papers)!
➥ [Research Highlight] The work has been introduced in KAIST School of Computing, Aug 2022.

[Our artwork] titled "the sneakers universe" was accepted for the AI x ART contest exhibit held by National Science Museum, Dec 2021 ➥ please see an IBS News article (in Korean).

[One abstract] titled "The Shape of Design History: Exploring Evolution of Sneakers Design at Scale Using Neural Embedding" was presented at IC2S2'21, Jul 2021.
➥ If you interested in this abstract please check the presentation video [here].

[One paper] titled "Neural Embedding of Sneaker Designs over 23 Years" was presented at KCC'21 (Korea Computer Congress), Jun 2020.

Close Window

More & More

[An extended abstract] to prpose approaches to successful entry of the ride-sharing service for startups has been presented at CHI'17, May 2017.

More exciting projects will come out shortly! Please stay tuned :)

Close Window