Soojung Yang

Research

Modeling protein dynamics and conformational ensembles with molecular simulations, foundation models, and experimental data. 

2025

Representations of Matter Across Scientific Foundation Models

Sathya Edamakada1, Soojung Yang1*, Rafael Gómez-Bombarelli*

NeurIPS AI4Mat Workshop, 1co-first, *co-corresponding

GAP: Guided Diffusion for A Priori Transition State Sampling

Hyukjun Lim1, Soojung Yang1*, Lucas Pinede, Miguel Steiner, Yuanqi Du, Rafael Gómez-Bombarelli*

NeurIPS FPI Workshop, 1co-first, *co-corresponding

When Do LLMs Improve Bayesian Optimization? A Systematic Comparison Across Molecular and Protein Design

Mattias Akke, Soojung Yang*, Jurgis Ruza, Jinyeop Song, Elton Pan, Rafael Gómez-Bombarelli*

NeurIPS AI4Science Workshop, *co-corresponding

Unifying Force Prediction and Molecular Conformation Generation Through Representation Alignment

Lucas Pinede1, Soojung Yang1, Juno Nam, Rafael Gómez-Bombarelli

ICML GenBio Workshop, 1co-first 

👉 Training of diffusion model-based Boltzmann emulator can be accelerated via alignment with pretrained MLIPs. 

Learning Collective Variables from Time-lagged Generation

Seonghyun Park, Kiyoung Seong, Soojung Yang, Rafael Gómez-Bombarelli, Sungsoo Ahn

ICML GenBio Workshop

Scalable emulation of protein equilibrium ensembles with generative deep learning

Sarah Lewis1, Tim Hempel1, José Jiménez Luna1, Michael Gastegger1, Yu Xie1, Andrew Y. K. Foong1, Victor García Satorras1, Osama Abdin1, Bastiaan S. Veeling1, Iryna Zaporozhets, Yaoyi Chen, Soojung Yang, Arne Schneuing, Jigyasa Nigam, Federico Barbero, Vincent Stimper, Andrew Campbell, Jason Yim, Marten Lienen, Yu Shi, Shuxin Zheng, Hannes Schulz, Usman Munir, Cecilia Clementi, Frank Noé

Science

👉 I developed PPFT (Property Prediction Fine-Tuning), a method to fine-tune a diffusion model with ensemble statistics. 

Transferable Learning of Reaction Pathways from Geometric Priors

Juno Nam, Miguel Steiner, Max Misterka, Soojung Yang, Avni Singhal, Rafael Gómez-Bombarelli

arXiv

2024

Probing the Embedding Space of Protein Foundation Models through Intrinsic Dimension Analysis

Soojung Yang, Juno Nam, Tynan Perez, Jinyeop Song, Xiaochen Du, Rafael Gómez-Bombarelli

NeurIPS AIDrugX Workshop

👉 Protein foundation models are aligned with each other, across modalities and architectures. 

Flow Matching for Accelerated Simulation of Atomic Transport in Materials

Juno Nam, Sulin Liu, Gavin Winter, KyuJung Jun, Soojung Yang, Rafael Gómez-Bombarelli

arXiv

Learning Collective Variables for Protein Folding with Labeled Data Augmentation through Geodesic Interpolation

Soojung Yang1, Juno Nam1, Johannes C. B. Dietschreit, Rafael Gómez-Bombarelli

JCTC, 1co-first 

👉 Geodesic interpolations can serve as a prior for CV for accelerated MD simulations.

2023

Regularized indirect learning improves phage display ligand discovery

Joseph S. Brown, Yitong Tseo, Michael A. Lee, Jeffrey Y.-K. Wong, Soojung Yang, Yehlin Cho, Chae Rin Kim, Andrei Loas, Ratmir Derda, Rafael Gómez-Bombarelli, Bradley L. Pentelute

ChemRxiv 

Chemically Transferable Generative Backmapping of Coarse-Grained Proteins

Soojung Yang, Rafael Gómez-Bombarelli

ICML

👉 Protein local fluctuations can be transferably learned from a structural ensemble database. 

2022

PIGNet: A physics-informed deep learning model toward generalized drug-target interaction predictions

Seokhyun Moon1, Wonho Zhung1, Soojung Yang1, Jaechang Lim, Woo Youn Kim

Chemical Science, 1co-first

2021

Hit and Lead Discovery with Explorative RL and Fragment-based Molecule Generation

Soojung Yang, Doyeong Hwang, Seul Lee, Seongok Ryu, Sung Ju Hwang

NeurIPS

2020

Comprehensive Study on Molecular Supervised Learning with Graph Neural Networks

Doyeong Hwang,  Soojung Yang , Yongchan Kwon , Kyung Hoon Lee, Grace Lee, Hanseok Jo, Seyeol Yoon, Seongok Ryu 

J Chem Inf Model

A comprehensive study on the prediction reliability of graph neural networks for virtual screening

Soojung Yang, Kyung Hoon Lee, Seongok Ryu

arXiv 

Community Service

• Program chair and organizer of the integrating Generative and Experimental platforms for bioMolecular design (GEM) at ICLR 2024 in Vienna, Austria, and ICLR 2025 in Singapore.  

Invited Talks & Poster Presentations

Transferable temperature-dependent Boltzmann Emulator 

• SSRL/LCLS Users' Meeting at SLAC National Laboratory: Building a General Inference Engine for Molecular Structures and Ensembles (Sep 2025)

Learning Collective Variables for Protein Folding with Labeled Data Augmentation through Geodesic Interpolation

• MIT Computational & Systems Biology Student Seminar (Feb 2024)

• IMSI workshop Learning Collective Variables and Coarse Grained Models (Apr 2024)

• MoML Best Paper Award (Nov 2024)

• Lab seminar invites: EPFL Michele Ceriotti Lab (Aug 2024), HMS Debora Marks Lab (Mar 2024)

Chemically Transferable Generative Backmapping of Coarse-Grained Proteins

• Learning on Graphs and Geometry (LoG) Reading Group (Apr 2023)

• CECAM/Psi-k Conference “Bridging length scales with machine learning: from wavefunctions to thermodynamics” Invited Talk/Panel Discussion (May 2023) 

• ICML Poster (Jul 2023)

• Boston Protein Design and Modeling Club Seminar (Dec 2023)

• Lab seminar invites: POSTECH Sungsoo Ahn Lab (Apr 2023), Flagship Pioneering  (May 2023) 

Hit and Lead Discovery with Explorative RL and Fragment-based Molecular Generation

• KAIST AI Student Colloquium (Oct 2021)

• NeurIPS Poster (Dec 2021)

Experience

Microsoft Research AI4Science

Research Intern (Mentor: Yu Xie, PI: Frank Noe)

AITRICS

ML Researcher

ACE (Advanced Computational Engine) Team, Dept. of Chemistry, KAIST 

Undergraduate Researcher (supervised by Prof. Woo Youn Kim)

Wellman Center of Photomedicine, MGH

Research Intern (supervised by Prof. Walfre Franco)

Selective Honors

• Takeda Fellowship, 2023-2024

• D.E.Shaw Research Graduate Women's Fellowship, 2023

• Ilju Foundation Scholarship for Ph.D Studies, 2021-2025