I am good at presentation and  full of patience. My research interest includes data science and deep neural network compression. When pursuing master’s degree, I have been a teaching assistant for calculus and a research assistant who focused on deep neural network compression and social network analysis.

Research 

Deep Neural Networks Compression

  • Implemented compression algorithms with the Pytorch framework.
  • Applied linearization to reduce the parameters in convolution layers and fully-connected layers.
  • Reduced at least 50% storage usage.
  • Increased 0.3% top-1 accuracy in ImageNet classification task.
  • Comprehensively analyzed the efficiency of the proposed compression scheme by calculus.
  • Outperformed the state-of-the-art algorithms published in IEEE journals (IEEE TNNLS).

Link Prediction for Social Recommendation

  • Utilized the regression model to predict hidden links between nodes in a social network.
  • Extracted features from the proximity of edges on a social network.
  • Defined criterion to seek hidden connections on the network.

College Credit Checker (project for NTUST)

  • Applied pandas to deal with excel raw data.
  • Designed a system to check student credits for graduation rule.
  • Mined the students’ interests from their course grades.

Education

2019 - 2021

National Taiwan University of Science and Technology

M.S. in Electronics and Computer Engineering

Research interests: Machine learning, Social network analysis, Mathematical optimization

2015-2019

National Taiwan University of Science and Technology

B.S. in Electronics and Computer Engineering

Publications

Journal:

  • H.-Y. Hsieh, S.-A. Huang*, J.-S. Leu, “Implementing a Real-Time Image Captioning Service for Scene Identification Using Embedded System,” Multimedia Tools and Applications (MTAP), Jan. 2021. https://doi.org/10.1007/s11042-020-10292-y
  • Y. -R. Chen, J. -S. Leu*, S. -A. Huang, J. -T. Wang and J. -I. Takada, "Predicting Default Risk on Peer-to-Peer Lending Imbalanced Datasets," in IEEE Access, vol. 9, pp. 73103-73109, 2021, doi: 10.1109/ACCESS.2021.3079701.
  • S. -A. Huang and J. -S. Leu*, “Piecewise Linearization with Dense Matrix Encoding for Neural Networks Parameters to Achieve Model Compression” IEEE Transactions on Neural Networks and Learning Systems. (under review)

Conference:

  • H.-Y. Hsieh, J.-S Leu* and S.-A. Huang, "Implementing a Real-Time Image Captioning Service for Scene Identification Using Embedded System," 2019 16th Annual IEEE International Conference on Sensing, Communication, and Networking (SECON), Boston, MA, USA, 2019, pp. 1-2, doi: 10.1109/SAHCN.2019.8824961.
Powered by CakeResumePowered by CakeResume