Doyoung Kim (κΉ€λ„μ˜)

My name is Doyoung Kim, and I am an incoming PhD student at NYU CS, advised by Prof. Sherry Yang.

My research interests is about general intelligence in langauge and robotics.

Currently I am a MS student at KAIST AI, advised by Prof. Minjoon Seo. Before studying AI, I completed my BS in Mathematics & Computer Science (double major) at KAIST.

Email  /  Google Scholar  /  X  /  Github  /  LinkedIn  /  CV

profile photo
Publications

Please see my Semantic Scholar or Google Scholar profiles for the full list.

* denotes equal contribution.
How language models extrapolate outside the training data: A case study in Textualized Gridworld
Doyoung Kim, Jongwon Lee, Jinho Park, Minjoon Seo
Neurips 2024 Compositional Learning Workshop
[paper] [blog]

While humans can learn complex reasoning from few examples, AI struggles to generalize beyond its training. We enable language models to generate "cognitive maps" - tree-structured expansions of future states - before planning. In maze-solving tasks, this cognitive mapping approach proves to be the only effective method for helping language models extrapolate their planning abilities to larger, unseen mazes.

Self-Explore: Enhancing Mathematical Reasoning in Language Models with Fine-grained Rewards
Hyeonbin Hwang, Doyoung Kim, Seungone Kim, Seonghyeon Ye, Minjoon Seo
EMNLP 2024 Findings
[paper]

We propose a self-training method that helps LLMs identify their first incorrect reasoning step ("pit") and use it as a reward signal. Through preference optimization, this method enables LLMs to improve their reasoning process, leading to enhanced mathematical performance.

Semiparametric Token-Sequence Co-Supervision
Hyunji Lee*, Doyoung Kim*, Jihoon Jun, Sejune Joo, Joel Jang, Kyoung-Woon On, Minjoon Seo
ACL 2024
[paper]

We introduce a semiparametric model superposing two embedding spaces: parametric token embeddings and nonparametric sequence embeddings. The model is co-trained using weighted cross-entropy loss for language modeling and InfoNCE loss for sequence retrieval to enable generation with citations.

FLASK: Fine-grained Language Model Evaluation based on Alignment Skill Sets
Seonghyeon Ye*, Doyoung Kim*, Sungdong Kim, Hyeonbin Hwang, Seungone Kim, James Thorne, Juho Kim, Minjoon Seo
ICLR 2024 Spotlight
[paper]

We propose a fine-grained evaluation framework for generative language models based on 12 alignment skill sets, which show a strong correlation between model-based and human-based evaluations.

The CoT Collection: Improving Zero-shot and Few-shot Learning of Language Models via Chain-of-Thought Fine-Tuning
Seungone Kim, Sejune Joo, Doyoung Kim, Joel Jang, Seonghyeon Ye, Jamin Shin, Minjoon Seo
EMNLP 2023
[paper]

We introduce a new instruction-tuning dataset called the COT COLLECTION dataset, containing 1.84 million rationales across 1,060 tasks. These rationales were extracted from the FLAN Collection using OpenAI Codex with in-context learning (ICL). We fine-tune Flan-T5 (3B & 11B) with the COT COLLECTION to show both zero-shot and few-shot improvements.


Show All Publications
Projects

* denotes equal contribution.
Selfee: Iterative self-revising llm empowered by self-feedback generation
Seonghyeon Ye*, Yongrae Jo*, Doyoung Kim*, Sungdong Kim, Hyeonbin Hwang, Minjoon Seo

[blog]


Website design from Jon Barron