Juntae (Jun-Tae) Lee (이준태)
About me
Currently, I am working as a Staff Researcher in Qualcomm AI Research. My research has been lied in computer vision and machine learning, especially audio-visual fusion, few-shot learning, personalization, efficient transfer learning, class-incremental learning. In current, I am focusing on researching large language model (LLM) and vision foundation models (VFM) for on-device.
I received the B.S. and Ph.D. degrees in the School of Electrical Engineering from Korea University in Feb. 2013 and Aug. 2019, respectively. I was a Staff Engineer in Visual S/W R&D Group, Samsung Electronics Mobile, Korea (Sep. 2019 - Apr. 2020).
Email: jtlee@mcl.korea.ac.kr
Tel: +82-10-2066-5104
Research Interest
Research Area: Machine Learning, Large Language Models, Computer Vision, Image Processing
Current Research Topics (in industrial area):
Large language models for edge devices
Vision-language model for edge devices
Previous Research Topics
Action recognition: few-shot, multi-modality (audio-visual), class-incremental learning
Few-shot learning
Transfer learning
Personalized deep networks
Personalized Image Aesthetics Estimation
Semantic Line Detection
Photographic Composition Classification
Education
2013 - 2019 Ph.D. Dept. of Electrical Engineering, Korea University (Advisor: Chang-Su Kim)
2009 - 2013 B.S. Dept. of Electrical Engineering, Korea University
Conference Publications
Jihwan Bang*, Juntae Lee*, Kyuhong Shim, Seunghan Yang, Simyung Chang, "Crayon: Customized On-Device LLM via Instant Adapter Blending and Edge-Server Hybrid Inference" accepted at ACL (main) 2024. *: equal contribution
Jun-Tae Lee, Mihir Jain, Sungrack Yun, "Few-shot common action localization via cross-attentional fusion of context and temporal dynamics," in Proc. IEEE ICCV 2023.
Jun-Tae Lee and Sungrack Yun, "Multi-scale temporal feature fusion for few-shot action recognition," in Proc. IEEE ICIP 2023.
Byeonggeun Kim*, Jun-Tae Lee*, Kyuhong Shim, Simyung Chang, "Task-agnostic open-set prototype for few-shot open-set recognition," in Proc. IEEE ICIP 2023. *: equal contribution
Byeonggeun Kim, Jun-Tae Lee*, Seunghan Yang, Simyung Chang, "Scalable Weight Reparametrization for Efficient Transfer Learning," in Proc. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2023 (Oral session) *:equal contribution
Jangho Kim*, Jun-Tae Lee*, Simyung Chang, "Variational On-the-Fly Personalization," in International Conference on Machine Learning (ICML), 2022 (Spotlight session) *: equal contribution
Byeonggeun Kim, Seunghan Yang, Jangho Kim, Hyunsin Park, Juntae Lee, Simyung Chang, "Domain Generalization with Relaxed Instance Frequency-wise Normalization for Multi-device Acoustic Scene Classification," in Proc. INTERSPEECH, 2022.
Jun-Tae Lee, Hyunsin Park, Sungrack Yun, Simyung Chang, "Multi-Head Modularization to Leverage Generalization Capability in Multi-Modal Network," in Proc. Thirty-Sixth AAAI Conference on Artificial Intelligence (AAAI), 2022
Jun-Tae Lee, Mihir Jain, Sungrack Yun, "Leaky Gated Cross-Attention for Weakly Supervised Multi-Modal Temporal Action Localization," in Proc. IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2022
Hanul Kim, Mihir Jain, Jun-Tae Lee, Sungrack Yun, Faith Porikli, "Efficient Action Recognition via Dynamic Knowledge Propagation," in Proc. IEEE International Conference on Computer Vision (ICCV), 2021
Jun-Tae Lee, Mihir Jain, Hyoungwoo Park, Sungrack Yun, "Cross-Attentional Audio-Visual Fusion for Weakly-Supervised Action Localization," in Proc. 9th International Conference on Learning Representation (ICLR), 2021
Dongkwon Jin, Jun-Tae Lee, and Chang-Su Kim, "Semantic line detection using mirror attention and comparative ranking and matching," in Proc. 16th European Conference on Computer Vision (ECCV), 2020
Jun-Tae Lee, and Chang-Su Kim, “Image Aesthetic Assessment Based on Pairwise Comparison – A Unified Approach to Score Regression, Binary Classification, and Personalization,” in Proc. IEEE International Conference on Computer Vision (ICCV), 2019
Jun-Tae Lee, Han-Ul Kim, Chul Lee, and Chang-Su Kim, “Semantic line detection and its applications,” in Proc. IEEE International Conference on Computer Vision (ICCV), 2017.
Keunsoo Ko, Jun-Tae Lee, and Chang-Su Kim, “PAC-Net: Pairwise aesthetic comparison network for image aesthetic assessment,” in Proc. IEEE International Conference on Image Processing (ICIP), Athens, Greece, Oct. 2018.
Jun-Tae Lee, Chulwoo Lee, Jae-Young Sim, and Chang-Su Kim, “Depth-guided adaptive contrast enhancement using 2D histograms,” in Proc. IEEE International Conference on Image Processing (ICIP), Paris, France, Oct. 2014.
Jun-Tae Lee, Kyung-Rae Kim, Won-Dong Jang, and Chang-Su Kim, “Near-duplicate video clustering using multiple complementary video signatures,” in Proc. Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Hong-Kong, Dec. 2015.
Jun-Tae Lee, Jae-Kyun Ahn, Chang-Su Kim, “Stitching of heterogeneous images using depth information,” in Proc. Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Kaohsiung, Taiwan, Nov. 2013.
Journal Publications
Jun-Tae Lee, Han-Ul Kim, Chul Lee, and Chang-Su Kim, “Photographic composition classification and dominant geometric element detection for outdoor scenes,” Journal of Visual Communication and Image Representation, vol. 55, no. 1, pp. 91–105, Aug. 2018.
Jun-Tae Lee, Chul Lee, and Chang-Su Kim, “Property-Specific Aesthetic Assessment With Unsupervised Aesthetic Property Discovery,” IEEE ACCESS, vol. 7, no. 1, pp. 114349–114362, Aug. 2019.