Tae-Hoon Kim

Abstract The Transformer architecture has revolutionized the field of deep learning over the past several years in diverse areas, including natural language processing, code generation, image recognition, and time-series forecasting. We propose to apply Zamir et al.'s efficient transformer to perform deconvolution and denoising to enhance astronomical images. We conducted experiments using pairs of high-quality images and their degraded versions, and our deep learning model demonstrates exceptio

Abstract Traditional weak-lensing mass reconstruction techniques suffer from various artifacts, including noise amplification and the mass-sheet degeneracy. In S. E. Hong et al., we demonstrated that many of these pitfalls of traditional mass reconstruction can be mitigated using a deep learning approach based on a convolutional neural network (CNN). In this paper, we present our improvements and report on the detailed performance of our CNN algorithm applied to next-generation wide-field (WF) o

To understand the impact of new pricing structure on residential electricity demands, we need a baseline model that captures every factor other than the new price. The standard baseline is a randomized control group, however, a good control group is hard to design. This motivates us to devlop data-driven approaches. We explored many techniques and designed a strategy, named LTAP, that could predict the hourly usage years ahead. The key challenge in this process is that the daily cycle of electri

This study aims to examine research trends on international students in South Korea from 2006 onward and to identify imbalances or gaps in research methods and topics, thereby suggesting directions for future studies. The analysis covers 312 studies on international students published in Korea Citation Index (KCI)-indexed journals between 2006 and 2025. The findings indicate rapid quantitative growth in research on international students over this period, with research topics predominantly focus

A novel microfluidic device is described that can detect the chemotactic response of motile bacterial cells (Escherichia coli) that swim toward a preferred nutrient with high resolution by sorting and concentrating them. The device consists of the typical Y-shaped microchannels that have been widely used in chemotaxis studies to attract cells toward a high concentration and a concentrator array integrated with arrow head-shaped ratchet structures beside the main microchannel to trap and accumula

Slogans play a crucial role in building the brand's identity of the firm. A slogan is expected to reflect firm's vision and the brand's value propositions in memorable and likeable ways. Automating the generation of slogans with such characteristics is challenging. Previous studies developed and tested slogan generation with syntactic control and summarization models which are not capable of generating distinctive slogans. We introduce a novel approach that leverages pre-trained transformer T5 m

Manipulation of micron-scale silicon particles has been investigated with optical tweezers implemented using a two-dimensional scanning trap driven with acousto-optic modulators. Spheres of latex, Poly(methyl methacrylate) (PMMA), silica, and silver-coated PMMA have been utilized to calibrate transverse trapping forces. The goal of this work is to non-invasively manipulate 10-20μm silicon-based devices in and around cells.

In this study, we propose an ultra-fast responsive PC based volatile organic compounds (VOCs) sensor using a nanoscale easy tear process inspired by commercially available `easy tear package'. Colloidal crystal-polydimethylsiloxane (PDMS) composite can be realized through nanoscale tear propagation along the interface between the outer surface of crystallized nanoparticles and bulk PDMS. Not only cuboid but also dome shaped colloidal crystal-PDMS composite is successfully obtained using the asse