* Researched about target based small molecule generative models for de novo design - SMILES based de novo generation - Graph based de novo generation* Researched molecule properties prediction model using Graphormer - Added new algorithms* Special lecture in Lectures on AI-driven Drug Discovery (LAIDD) program - Title: Deep Learning Based Molecular Generation (https://www.laidd.org/course/123) - GitHub: https://github.com/ilguyi/LAIDD-molecular-generation

* Researched target based small molecule generative models for de novo design - Genertive models: used autoregressive models, VAE, GAN - NLP models: transformer based (Bert, GPT) models - Graph based models and reinforcement learning* Protein Design - Researched new model that can generate sequences given backbone structure - Used graph based model and transformer based model* Development of small molecule property prediction model - Applied molecule's 3D structure information to Bert based model* Drug-Target Interaction - Developed deep learning based DTI models - Developed sequence and structure based models

* Researched generative models* Lecture on generative models* GTA Lab leader in Modulabs - Researched music generation - Researched music source separation

* Computer Vision based on Deep Learning - Image classification - Object detection - Activity classification - Body part detection - Image captioning - Generative Adversarial Networks