lucainflo

Importance Notification

Starting in September 2024, I began my research as an Assistant Professor at Sungshin Women's University. A new website for SAIL(Scalable Artificial Intelligence Lab) will be launched soon. Please contact me at yhmoon_at_sungshin_dot_ac_dot_kr.

Join Our Lab (UST at ETRI School)

• (New Positions) We are currenlty looking for Ms/PhD students and Post-Doctoral Researchers. Full scholarships for graduate students are guaranteed and we also support powerful training GPU machines and experimental environments. Research subjects and directions will be carefully advised and fully discussed in our Lab. Please see [Contact] at the bottom."

• Youtube Video
• Admission (Korean)
• Admission (English)
• Apply UST!

Current Activities

• (Professional Service) "Serving as a reviewer for EMNLP 2024 (July 2024)".

• (Edge AI Workshop 2024) " Edge AI Workshop associated with KCC 2024 (June 28 2024)".

• (Edge AI Summer Internship 2024) " Edge AI Summer Intership Program 2024 associated with ETRI (July 1 2024 ~ August 31 2024)".

• (Professional Service) "Serving as a reviewer for Neurocomputing Journal (May 2024)".

• (Professional Service) "Serving as a reviewer for ICPR 2024 (May 2024)".

• (Professional Service) "Serving as a reviewer for ACL 2024 (Feb. 2024 ~ March 2024)".

• (New Student) "Sungyoun Lim starts his Ph.D. program under the supervision of Prof. Yong-Hyuk Moon. (March 1, 2024)".

• (Professional Service) "Serving as a reviewer for CVPR 2024 (Dec. 1, 2023 ~ Feb. 28, 2024)".

• (Edge AI Workshop 2023) "Edge AI Workshop associated with KCC 2023 (June. 2024)".

• (Edge AI Winter Internship 2023) "Edge AI Winter Intership Program 2023 associated with ETRI (Jan. 1 2024 ~ Feb. 28 2024)".

• (The 3rd GIST-ETRI AI Workshop 2022) I am grateful to share my talk about "Multiscaled Edge AI Technology: Motivation, Research Trends, and Our Achievements," held in GIST, Gwangu City, July 28, 2022.

• (The 1st Visual Intelligence Workshop 2022) I shared my talk about "Edge AI Research Treads: Compression, Neural Architecture Search, Multiscaling," associated with KCC(Korea Computer Congress) 2022 held in Jeju Island, June 30 2022."

• (Edge AI Workshop 2022) Talks about nn-multiscaling, nn-disttraining, nn-runtime, edgeai-applications and cloud-edge framework are shared and related topics are freely discussed with our edge.ai partners, which are KETI, KAIST, Nota, FarmConnect, and Daejeon City."

• (Edge AI Summer Internship) "Edge AI Summer Intership Program 2022 associated with ETRI (July 1 ~ August 31)".

• (Best Paper Awards) Eunji Kim and Seongwoo Kim (Winter Internship 2022) won the Best Paper Awards in ASK 2022 organized by KIPS. - "A Docker-based Evaluation Program for Model Inference Performance on Heterogeneous Edge Environments" (Mentor: Jongryul Lee, Advisor: Yong-Hyuk Moon).

• (Paper Awards) Eunkyeoung Kim (Winter Intenship 2022) won the Paper Awards in ASK 2022 organized by KIPS. - "Instance Segmentation Based Tomato Pests Disease Detection for Feasibility Evaluation" (Mentor: Junyong Park, Advisor: Yong-Hyuk Moon).

Research

I'm interested in solving complex problems with staggeringly high dimensional data. Recently, my work explores subjects in computer vision and deep learning, more specifically concerning AutoML, Meta Learning, Model Scaling, and Distributed Learning.

# Our Journey starting from the ML pipeline
~ $ Searching near-optimal backbone arhcitectures
~ $ Adjusting hyperparameters and retraining
~ $ Multiscaling the deep nerual networks

# But way more beyond that for real-world analytics
~ $ Put more efforts for performance fine-tuning
~ $ Accelerating inference at edge devices
~ $ Deploy fine-tuned models on the ground!
~ $ Domain adaptation using distributed learning

Professor

• Yong-Hyuk Moon received the M.S. and Ph.D. degrees in Information and Communication Engineering from Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2006 and 2013, respectively. He also started serving as an Associate Professor in the Computer Software Department at the University of Science and Technology (UST), Daejeon, Korea, in September 2021. His particular research activities concern automated machine learning, deep learning, optimization theory, and distributed/edge computing. Academic achievements are as follows.

• Publication: 61 peer-reviewed research papers including SCI/SCIE/ESCI Journal (8), IEEE/ACM (19), LNCS/LNEE (4), Best/Selected/Invited (5), international conference papers (6), and domestic conference papers (19)

• TPC/Reviewer: IEEE TPDS, IEEE TSMC, IEEE JSAC, IEEE Wireless Magazine, IEEE Access, Journal of Supercomputing, Journal of Combinatorial Optimization, ETRI Journal, and lots of IEEE Conferences

• Patents/Standardization: 50 applications and 21 granted among them, and contributed to ITU-T (4) and ASTAP (8)

DNOL Members

• Currently, four AI researchers have been involved in the edge.ai project.
• Junyong Park (Task Lead, PhD Candidate) and Sungyoun Lim (Researcher, PhD Student) <!–
• Jong-Ryul Lee (Task Lead, PhD)and Yong-Ju Lee (PhD). –>

Projects Granted

(New) agent.ai

• (2022.04 ~ 2026.12 as Task Lead) More details will be provided soon.

edge.ai

• (2021.04 ~ 2024.12 as Project Lead) Development of Adaptive and Lightweight Edge-Collaborative Analysis Technology for Enabling Proactively Immediate Response and Rapid Learning (This work was supported by Institute of Information communications Technology Planning Evaluation (IITP) grant funded by the Korea government (MSIT)).

edge.analytics

• (2018.04 ~ 2020.12 as Task Lead) Development of Big Data Edge Analytics SW Technology for Load Balancing and Active Timely Response (This work was supported by Institute of Information communications Technology Planning Evaluation (IITP) grant funded by the Korea government (MSIT)).

New Publication

• (New: AI Top-Tier) Our new paper has been submitted to the 27th International Conference on Pattern Recognition (ICPR) 2024.

• (New: AI Top-Tier) Jongryul Lee and Yong-Hyuk Moon, "Revisiting Layer-level Residual Connections for Efficient Object Detection," accepted in the 20th edition of the IEEE International Conference on Advanced Video and Signal-Based Surveillance (AVSS) 2024.

• (New: AI Top-Tier) Jongryul Lee and Yong-Hyuk Moon, "Bespoke: A Block-Level Neural Network Optimization Framework for Low-Cost Deployment," in Association for the Advancement of Artificial Intelligence (AAAI) 2023.

• (New: AI Top-Tier) Jongryul Lee and Yong-Hyuk Moon, "Rethinking Group Fisher Pruning for Efficient Label-Free Network Compression," published in The 33rd British Machine Vision Association and Society for Pattern Recognition (BMVC) 2022, London, UK, November 21-24, 2022 (See https://bmvc2022.mpi-inf.mpg.de/693/).

• (IEEE Journal, SCIE-indexed, IF 3.367) Jong-Yeon Park, Yong-Hyuk Moon, Wonil Lee, Sung-Hyun Kim, and Kouichi Sakurai, "A Survey of Polynomial Multiplication With RSA-ECC Coprocessors and Implementations of NIST PQC Round3 KEM Algorithms in Exynos2100," published in IEEE ACCESS Volume 10, 2022.

• (AI Top-Tier, H5-Index 112) Jong-Ryul Lee, Yong-Ju Lee, and Yong-Hyuk Moon, "Block-wise Word Embedding Compression Revisited: Better Weighting and Structuring," in the proceedings of The 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP), November 7, 2021.

• (AI Top-Tier, H5-index 197) Jungyong Park, Yong-Hyuk Moon, and Yong-Ju Lee, "Efficient Approximation of Filters for High-Accuracy Binary Convolutional Neural Networks," European Conference on Computer Vision (ECCV) 2020 (LNCS 12539), v.12539, pp.71-84.

• (IEEE, SCOPUS-indexd) Yong-Hyuk Moon and Yong-Ju Lee, "A Practical Dense-to-Sparse Learning for a Manageable Object Detector on the Ground," to appear in the proceedings of the 12th International Conference on ICT Convergence (ICTC) 2021, October 20-22, 2021.

• (IEEE, SCOPUS-indexd) Jong-Ryul Lee and Yong-Hyuk Moon, "An Empirical Study on Channel Pruning through Residual Connections," to appear in the proceedings of the 12th International Conference on ICT Convergence (ICTC) 2021, October 20-22, 2021.

• (IEEE, SCOPUS-indexd) Jungyong Park, Jong-Ryul Lee, and Yong-Hyuk Moon, "Improved Early Exiting Activation to Accelerate Edge Inference," to appear in the proceedings of the 12th International Conference on ICT Convergence (ICTC) 2021, October 20-22, 2021.

• (IEEE, SCOPUS-indexd) Jiyeon Kim, Yong-Ju Lee, and Yong-Hyuk Moon, "Enhancing the Data Regularization Effect with Randomly Combined Features for Object Detection," to appear in the proceedings of the 12th International Conference on ICT Convergence (ICTC) 2021, October 20-22, 2021.

Internship Programs

• (2024.07.01~2024.08.31) Edge AI Summer Intership Program 2024 associated with ETRI (Interns: Coming Up).

• (2024.01.01~2024.02.29) Edge AI Winter Intership Program 2024 associated with ETRI (Interns: Sungyoun Lim).

• (2023.01.01~2023.02.28) Edge AI Summer Intership Program 2023 associated with ETRI (Interns: Yeojin Lee).

• (2022.07.01~2022.08.31) Edge AI Summer Intership Program 2022 associated with ETRI (Interns: Seungsang Kwon).

• (2022.01.03~2022.02.25) Edge AI Winter Intership Program 2022 associated with UST (Interns: Jungmo Kim and Seongwoo Kim).

• (2022.01.01~2022.02.28) Edge AI Winter Intership Program 2022 associated with ETRI (Interns: Eunji Kim, Eunkyeoung Kim, and Seungwoo Roh).

Contact

• Contact and we are looking for grduate students (MS/PhD) and Post Drs.
• For more details, see our [Project Page] (work-in-progress).