SimBa: Simplicity Bias for Scaling Up Parameters in Deep Reinforcement Learning
Hojoon Lee*, Dongyoon Hwang*, Donghu Kim, Hyunseung Kim, Jun Jet Tai, Kaushik Subramanian, Peter R. Wurman, Jaegul Choo, Peter Stone, Takuma Seno.
ICLR'25 (Spotlight).
project page / paper

Designed network architectures that steer convergence toward simple functions which allows to scale up parameters in RL.

Do’s and Don’ts: Learning Desirable Skills with Instruction Videos
Hyunseung Kim, Byungkun Lee, Hojoon Lee, Dongyoon Hwang, Donghu Kim, Jaegul Choo
NeurIPS'24.
project page / paper

We present DoDont, an instruction-based skill discovery algorithm designed to combine human intention with unsupervised skill discovery. DoDont learns diverse behvaiors while following the behaviors in "do" videos while avoiding the behaviors in "don't" videos.

Slow and Steady Wins the Race: Maintaining Plasticity with Hare and Tortoise Networks
Hojoon Lee, Hyeonseo Cho, Hyunseung Kim, Donghu Kim, Dugki Min, Jaegul Choo, Clare Lyle
ICML'24.
paper / code

To allow the network to continually adapt and generalize, we introduce Hare and Tortoise architecture, inspired by the complementary learning system of the human brain.

A Simple Convolution INjector for Vision Transformer: Towards Effective Adaptation in Visuo-Motor Control
Dongyoon Hwang*, Byungkun Lee*, Hojoon Lee, Hyunseung Kim, Jaegul Choo
ICML'24.
paper

We introduce CoIn, a lightweight and simple add-on module, which effectively adapts pretrained Vision Transformers for visuo-motor control.

DISCO-DANCE: Learning to Discover Skills through Guidance
Hyunseung Kim*, Byungkun Lee*, Hojoon Lee, Dongyoon Hwang, Sejik Park, Kyushik Min, Jaegul Choo
NeurIPS'23.
project page / paper / code / poster

We introduce DISCO-DANCE, a skill discovery algorithm focused on learning diverse, task-agnostic behaviors. DISCO-DANCE addresses the common limitation of exploration in skill discovery algorithms through explicit guidance.

PLASTIC: Improving Input and Label Plasticity for Sample Efficient Reinforcement Learning
Hojoon Lee*, Hanseul Cho*, Hyunseung Kim*, Daehoon Gwak, Joonkee Kim, Jaegul Choo SeYoung Yun, Chulhee Yun,
NeurIPS'23.
paper / code / slide / poster

For sample-efficient RL, the agent needs to quickly adapt to various inputs (input plasticity) and outputs (label plasticity). We present PLASTIC, which maintains both input and label plasticity by identifying smooth local minima and preserving gradient flow.

DraftRec: Personalized Draft Recommendation for Winning in MOBA Games
Hojoon Lee*, Dongyoon Hwang*, Hyunseung Kim, Byungkun Lee, Jaegul Choo
WWW'22.
paper / code / poster / Bibtex

We gathered data from 280,000 matches played by the top 0.3% rank players in Korea for League of Legends. From this, we developed DraftRec, a personalized champion recommendation system aimed at maximizing players' win rates.