Juyeon Heo

News

[2023. 12] One paper, "Use Perturbations when Learning from Explanations", is accepted to NeurIPS 2023. (Equal first)

[2023. 10] One paper, "Physics‐based Decoding Improves MRF", is accepted to MICCAI 2023.  (Equal first)

My work was featured as the main article of the MICCAI Daily magazine!

[2023. 07] One paper, "Leveraging Task Structures for Improved Identifiability in Neural Network Representations", is accepted to ICML 2023 workshop

[2023. 05] One paper, "GeValDi: Generative Validation of Discriminative Models", is accepted to ICLR 2023 workshop. 

[2023. 05] One paper, "Robust Explanation Constraints for Neural Networks", is accepted to ICLR 2023.  (Equal first)

[2022. 12] One paper, "Towards More Robust Interpretation via Local Gradient Alignment", is accepted to AAAI 2023.

[2021. 10. 01] Start pursuing an Engineering Ph.D. in Cambridge, UK.

[2020. 08. 23] Graduation on master course in SKKU.

[2020. 03. 01] Visiting student in TU Berlin with Prof. Klaus-Robert Müller.

[2019. 09. 03] One paper, "Fooling Neural Network interpretations via adversarial model manipulation", is accepted to NeurIPS 2019.  (Equal first)

[2019. 10. 23] One paper, "Estimating PM2.5 concentration of the conterminous United States via interpretable convolutional neural networks", is accepted to Environmental Pollution.

[2019. 10. 30] One paper, "Interpreting machine learning models in neuroimaging: Towards a unified framework", is accepted to Nature Protocols.

Education

PhD [2021.10 ~ ]

• Engineering in Cambridge, UK

• Supervisor: Adrian Weller

• Cambridge Trust Scholarship (Schlumberger Cambridge International Scholarship)

MS [2018.09 ~ 2020.08]

• Electrical and Computer Engineering in SKKU

• GPA: 4.26 / 4.5

BS [2014.03 ~ 2018.08]

• Global Business Administration in SKKU

• GPA: 4.03 / 4.5

Publication

• Physics‐based Decoding Improves MRF

  • MICCAI 2023

• [C1] Robust Explanation Constraints for Neural Networks

  • ICLR 2023

• [C1] Towards More Robust Interpretation via Local Gradient Alignment

  • AAAI 2023

• [C1] Fooling Neural Network interpretations via adversarial model manipulation

  • • Juyeon Heo, Sunghwan Joo, and Taesup Moon

    • Proceedings of Neural Information Processing Systems (NeurIPS), December 2019 (Acceptance rate: 21.1%)

    • Code and video on https://github.com/rmrisforbidden/Fooling_Neural_Network-Interpretations

• [Journal] Interpreting machine learning models in neuroimaging: Towards a unified framework

• • Lada Kohoutová, Juyeon Heo, Sungmin Cha, Sungwoo Lee, Taesup Moon, Tor D. Wager, and Choong-Wan Woo

  • Nature Protocols

• [Journal] Estimating PM2.5 concentration of the conterminous United States via interpretable convolutional neural networks

• • Yongbee Park, ByungjoonKwon, Juyeon Heo, Xuefei Hu, Yang Liu, and Taesup Moon

  • Environmental Pollution

Experiences

2021.10 ~ : Pursuing an Engineering Ph.D, Cambridge, UK

2020.09 ~ 2021.08 :  Post master researcher, M.IND.LAB, SKKU

2020.03 ~ 2020.08 : Visiting student in TU Berlin (Prof. Klaus-Robert Müller)

2019.02 ~ 2019.07 : Teaching Assistant (class : Neural Networks), SKKU

2018.08 ~ 2018.12 : Teaching Assistant (class : Deep Neural Networks), SKKU

Scholarship

2021.10 ~ 2024.09 : Cambridge Trust Scholarship (Schlumberger Cambridge International Scholarship), Cambridge, UK

2018.08 ~ 2020.09 : Brain Korea (BK) Scholarship, SKKU

2018.08 ~ 2020.09 : Simsan Scholarship, SKKU

2018.03 - 2018.08 : Academic Excellence Scholarship, SKKU

2014.12 , 2017.12 : Dean's List of Business, SKKU

2015.03 - 2015.08 : Academic Excellence Scholarship, SKKU