I am a PhD student at Machine Learning and Artificial Intelligence (MLAI) lab in KAIST, working on large language models and safety alignment. I am fortunate to be advised by Professor Sung Ju Hwang and Juho Lee. I also closely collaborate with Kenji Kawaguchi. My research is graciously supported by the Apple Scholars in AI Fellowship. Here is my cv.

During my PhD study, I interned at Mila in 2024, where I had the opportunity to work with Yoshua Bengio, Minsu Kim, Moksh Jain, and Kolya Malkin. In 2023, I completed an internship at Apple Cambridge, working with Anders Johannsen and Jianpeng Cheng. In 2022, I did remote internship at NUS, working with Kenji Kawaguchi.

📖 Educations

• 2022.03 - present, PhD. in Artificial Intelligence. Korea Advanced Institute of Science and Technology.
• 2020.03 - 2022.02, M.S. in Artificial Intelligence. Korea Advanced Institute of Science and Technology.
• 2011.03 - 2018.02, B.A. in Library and Information Science. Yonsei University.

💻 Work Experience

• 2024.01 - 2024.06, Internship at Mila. Host: Yoshua Bengio.

• 2023.06 - 2023.09, Internship at Apple Cambridge. Host: Anders Johannsen.
• 2022.06 - 2022.09, Remote internship at NUS. Host: Kenji Kawaguchi.

📝 Publications

• SafeRoute: Adaptive Model Selection for Efficient and Accurate Safety Guardrails in Large Language Models

  [paper]
  Seanie Lee*, Dong Bok Lee*, Dominik Wagner, Minki Kang, Haebin Seong, Tobias Bocklet, Juho Lee and Sung Ju Hwang (*: equal contribution)
  ACL Findings 2025

• Personalized Fine-Tuning with Controllable Synthetic Speech from LLM-Generated Transcripts for Dysarthric Speech Recognition

  [paper]
  Dominik Wagner, Ilja Baumann, Natalie Engert, Seanie Lee, Elmar Nöth, Korbinian Riedhammer and Tobias Bocklet
  Interspeech 2025

• FedSVD: Adaptive Orthogonalization for Private Federated Learning with LoRA

  [paper]
  Seanie Lee*, Sangwoo Park*, Dong Bok Lee*, Dominik Wagner, Haebin Seong, Tobias Bocklet, Juho Lee and Sung Ju Hwang (*: equal contribution)
  Arxiv 2025

• Distilling LLM Agent into Small Models with Retrieval and Code Tools

  [paper]
  Minki Kang, Jongwon Jeong, Seanie Lee, Jaewoong Cho and Sung Ju Hwang
  Arxiv 2025

• Trajectory Balance with Asynchrony: Decoupling Exploration and Learning for Fast, Scalable LLM Post-Training

  [paper]
  Brian R. Bartoldson, Siddarth Venkatraman, James Diffenderfer, Moksh Jain, Tal Ben-Nun, Seanie Lee, Minsu Kim, Johan Obando-Ceron, Yoshua Bengio, Bhavya Kailkhura
  Arxiv 2025

• HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models

  [paper]
  Seanie Lee*, Haebin Seong*, Dong Bok Lee, Minki Kang, Xiaoyin Chen, Dominik Wagner, Yoshua Bengio, Juho Lee and Sung Ju Hwang (*: equal contribution)
  ICLR 2025

• Learning Diverse Attacks on Large Language Models for Robust Red-teaming and Safety Tuning

  [paper]
  Seanie Lee, Minsu Kim, Lynn Cherif, David Dobre, Juho Lee, Sung Ju Hwang, Kenji Kawaguchi, Gauthier Gidel, Yoshua Bengio, Nikolay Malkin and Moksh Jain
  ICLR 2025

• Calibrated Decision-Making through LLM-Assisted Retrieval

  [paper]
  Chaeyun Jang, Hyungi Lee, Seanie Lee and Juho Lee
  arXiv 2024

• Optimized Speculative Sampling for GPU Hardware Accelerators

  [paper]
  Dominik Wagner, Seanie Lee, Ilja Baumann, Philipp Seeberger, Korbinian Riedhammer and Tobias Bocklet
  EMNLP 2024

• Drug Discovery with Dynamic Goal-aware Fragments

  [paper]
  Seul Lee, Seanie Lee, Kenji Kawaguchi and Sung Ju Hwang
  ICML 2024

• Effective and Efficient Conversation Retrieval for Dialogue State Tracking with Implicit Text Summaries

  [paper]
  Seanie Lee, Jianpeng Cheng, Joris Driesen, Alexandru Coca and Anders Johannsen
  NAACL 2024

• Self-Supervised Dataset Distillation for Transfer Learning

  [paper]
  Dong Bok Lee*, Seanie Lee*, Joonho Ko, Kenji Kawaguchi, Juho Lee and Sung Ju Hwang (*: equal contribution)
  ICLR 2024

• DiffusionNAG: Task-guided Neural Architecture Generation with Diffusion Models

  [paper]
  Sohyun Ahn* Hayeon Lee*, Jaehyeong Jo, Seanie Lee and Sung Ju Hwang (*: equal contribution)
  ICLR 2024

• Knowledge-Augmented Reasoning Distillation for Small Language Models in Knowledge-Intensive Tasks

  [paper]
  Minki Kang, Seanie Lee, Jinheon Baek, Kenji Kawaguchi and Sung Ju Hwang
  NeurIPS 2023

• Scalable Set Encoding with Universal Mini-Batch Consistency and Unbiased Full Set Gradient Approximation

  [paper]
  Jeffrey Willette*, Seanie Lee*, Bruno Andreis, Kenji Kawaguchi, Juho Lee and Sung Ju Hwang (*: equal contribution)
  ICML 2023

• Margin-based Neural Network Watermarking

  [paper]
  Byungjoo Kim, Suyoung Lee, Seanie Lee, Sooel Son and Sung Ju Hwang
  ICML 2023

• Self-Distillation for Further Pre-training of Transformers

  [paper]
  Seanie Lee, Minki Kang, Juho Lee, Sung Ju Hwang and Kenji Kawaguchi
  ICLR 2023

• Self-Supervised Set Representation Learning for Unsupervised Meta-Learning

  [paper]
  Dong Bok Lee*, Seanie Lee*, Kenji Kawaguchi, Yunji Kim, Jihwan Bang, Jung-Woo Ha and Sung Ju Hwang (*: equal contribution)
  ICLR 2023

• Set-based Meta-Interpolation for Few-Task Meta-Learning

  [paper]
  Seanie Lee*, Bruno Andreis*, Kenji Kawaguchi, Juho Lee and Sung Ju Hwang (*: equal contribution)
  NeurIPS 2022

• On Divergence Measures for Bayesian Pseudocoresets

  [paper]
  Balhae Kim, Jungwon Choi, Seanie Lee, Yoonho Lee, Jung-Woo Ha and Juho Lee
  NeurIPS 2022

• Set Based Stochastic Subsampling

  [paper]
  Bruno Andreis, Seanie Lee, A. Tuan Nguyen, Juho Lee, Eunho Yang and Sung Ju Hwang
  ICML 2022

• Sequential Reptile: Inter-Task Gradient Alignment for Multilingual Learning

  [paper]
  Seanie Lee*, Hae Beom Lee*, Juho Lee and Sung Ju Hwang (*: equal contribution)
  ICLR 2022

• Learning to Perturb Word Embeddings for Out-of-distribution QA

  [paper]
  Seanie Lee*, Minki Kang*, Juho Lee and Sung Ju Hwang (*: equal contribution)
  ACL 2021

• Contrastive Learning with Adversarial Perturbations for Conditional Text Generation

  [paper]
  Seanie Lee*, Dong Bok Lee* and Sung Ju Hwang (*: equal contribution)
  ICLR 2021

• Meta-GMVAE: Mixture of Gaussian VAE for Unsupervised Meta-Learning

  [paper]
  Dong Bok Lee, Dongchan Min, Seanie Lee and Sung Ju Hwang
  ICLR 2021

• Generating Diverse and Consistent QA pairs from Contexts with Information-Maximizing Hierarchical Conditional VAEs

  [paper]
  Dong Bok Lee*, Seanie Lee*, WooTae Jeong, Donghwan Kim and Sung Ju Hwang (*: equal contribution)
  ACL 2020

• g2pM: A Neural Grapheme-to-Phoneme Conversion Package for Mandarin Chinese Based on a New Open Benchmark Dataset

  [paper]
  Kyubyong Park* and Seanie Lee* (*: equal contribution)
  INTERSPEECH 2020

🎖 Honors and Awards

• 2023.03 A 2023 recipient of the Apple Scholars in AI ML PhD fellowship.
• Google Travel Grant for NeurIPS 2022 from Google
• 2018.12 Silver Medal in NLP challenge.

💬 Invited Talks

• 2023.10, Tech. Talk, Technische Hochschule Nürnberg Georg Simon Ohm. Present Scalable Set Encoding with Universal Mini-Batch Consistency and Unbiased Full Set Gradient Approximation.
• 2023.05, Tech. Talk, Samsung SDS. Present Scalable Set Encoding with Universal Mini-Batch Consistency and Unbiased Full Set Gradient Approximation.
• 2021.12, Tech. Talk, NAVER corp. Present ACL 2020 paper.