Sooyoung Jang

Daeun Jung, Jaehyeok Jang, Sooyoung Jang et al.

Computed tomography (CT) and clinical numeric data are essential modalities for cancer evaluation, but building large-scale multimodal training datasets for developing medical foundation models remains challenging due to the structural complexity of multi-slice CT data and high cost of expert annotation. In this study, we propose MEDFORM, a multimodal pre-training strategy that guides CT image representation learning using complementary information from clinical data for medical foundation model development. MEDFORM efficiently processes CT slice through multiple instance learning (MIL) and adopts a dual pre-training strategy: first pretraining the CT slice feature extractor using SimCLR-based self-supervised learning, then aligning CT and clinical modalities through cross-modal contrastive learning. Our model was pre-trained on three different cancer types: lung cancer (141,171 slices), breast cancer (8,100 slices), colorectal cancer (10,393 slices). The experimental results demonstrated that this dual pre-training strategy improves cancer classification performance and maintains robust performance in few-shot learning scenarios. Code available at https://github.com/DigitalHealthcareLab/25MultiModalFoundationModel.git

Sooyoung Jang, Hyung-Il Kim

Encouraging exploration is a critical issue in deep reinforcement learning. We investigate the effect of initial entropy that significantly influences the exploration, especially at the earlier stage. Our main observations are as follows: 1) low initial entropy increases the probability of learning failure, and 2) this initial entropy is biased towards a low value that inhibits exploration. Inspired by the investigations, we devise entropy-aware model initialization, a simple yet powerful learning strategy for effective exploration. We show that the devised learning strategy significantly reduces learning failures and enhances performance, stability, and learning speed through experiments.

Jongmin Park, Sooyoung Jang, Younghoon Shin

In this study, reinforcement learning was applied to learning two-dimensional path planning including obstacle avoidance by unmanned aerial vehicle (UAV) in an indoor environment. The task assigned to the UAV was to reach the goal position in the shortest amount of time without colliding with any obstacles. Reinforcement learning was performed in a virtual environment created using Gazebo, a virtual environment simulator, to reduce the learning time and cost. Curriculum learning, which consists of two stages was performed for more efficient learning. As a result of learning with two reward models, the maximum goal rates achieved were 71.2% and 88.0%.