HSHyun Ah Song
Hyun Ah Song
*This is an algorithmically generated author record based on publicly available information.
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Hyun Ah Song has a total of 19 co-authors. Here are the top co-authors by number of shared publications:
- Christos Faloutsos
Carnegie Mellon University
4 shared publications - Nicholas Sidiropoulos
Department of Electrical and Computer Engineering, University of Virginia, Charlottesville, VA, USA
3 shared publications Christos Faloutsos
Carnegie Mellon University
2 shared publications- Evangelos E. Papalexakis
University of California, Riverside, Riverside, CA, USA
2 shared publications - Faisal M. Almutairi
University of Minnesota, Minneapolis, USA
2 shared publications - Fan Yang
School of Microelectronics, Fudan University
2 shared publications
Latest Publications
- Open Access
- 7 July 2023
- Published by Springer Nature in The VLDB Journal
- Vol. 33 (1) , 253
Correction to: TurboLift: fast accuracy lifting for historical data recovery
- 7 July 2023
- Published by Springer Nature in The VLDB Journal
- Vol. 33 (1) , 253
Open AccessCorrection to: TurboLift: fast accuracy lifting for historical data recovery
- 9 March 2020
- Published by Springer Nature in The VLDB Journal
- Vol. 29 (5) , 1129-1148
TurboLift: fast accuracy lifting for historical data recovery
- 9 March 2020
- Published by Springer Nature in The VLDB Journal
- Vol. 29 (5) , 1129-1148
TurboLift: fast accuracy lifting for historical data recovery
- 16 October 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 31 (12) , 2304-2318
Efficient and Distributed Generalized Canonical Correlation Analysis for Big Multiview Data
Generalized canonical correlation analysis (GCCA) integrates information from data samples that are acquired at multiple feature spaces (or ‘views’) to produce low-dimensional representations. Since the 1960s, GCCA has attracted much attention in machine learning and data mining. Despite these efforts, the existing GCCA algorithms have serious complexity issues. The memory and computational complexities of the existing algorithms usually grow as a quadratic and cubic function of the problem dimension (the number of samples / features), respectively -- e.g., handling views with $\approx 1,000$ features using such algorithms already occupies $\approx 10^6$ memory and the per-iteration complexity is $\approx 10^9$ flops. To circumvent such difficulties, we first propose a GCCA algorithm whose memory and computational costs scale linearly in the problem dimension and the number of nonzero data elements, respectively. Consequently, the proposed algorithm can easily handle very large sparse views whose sample and feature dimensions both exceed $\approx 100,000$ . Our second contribution lies in proposing two distributed algorithms for GCCA, which compute the canonical components of different views in parallel and thus can further reduce the runtime significantly if multiple cores are available. We provide detailed convergence analyses of the proposed algorithms. Judiciously designed synthetic and real-data experiments are employed to showcase the effectiveness of the proposed algorithms.- 16 October 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 31 (12) , 2304-2318
Efficient and Distributed Generalized Canonical Correlation Analysis for Big Multiview Data
Generalized canonical correlation analysis (GCCA) integrates information from data samples that are acquired at multiple feature spaces (or ‘views’) to produce low-dimensional representations. Since the 1960s, GCCA has attracted much attention in machine learning and data mining. Despite these efforts, the existing GCCA algorithms have serious complexity issues. The memory and computational complexities of the existing algorithms usually grow as a quadratic and cubic function of the problem dimension (the number of samples / features), respectively -- e.g., handling views with $\approx 1,000$ features using such algorithms already occupies $\approx 10^6$ memory and the per-iteration complexity is $\approx 10^9$ flops. To circumvent such difficulties, we first propose a GCCA algorithm whose memory and computational costs scale linearly in the problem dimension and the number of nonzero data elements, respectively. Consequently, the proposed algorithm can easily handle very large sparse views whose sample and feature dimensions both exceed $\approx 100,000$ . Our second contribution lies in proposing two distributed algorithms for GCCA, which compute the canonical components of different views in parallel and thus can further reduce the runtime significantly if multiple cores are available. We provide detailed convergence analyses of the proposed algorithms. Judiciously designed synthetic and real-data experiments are employed to showcase the effectiveness of the proposed algorithms.- 24 May 2018
- Published by Springer Nature in Knowledge and Information Systems
- Vol. 59 (3) , 497-522
TensorCast: forecasting and mining with coupled tensors
- 24 May 2018
- Published by Springer Nature in Knowledge and Information Systems
- Vol. 59 (3) , 497-522
TensorCast: forecasting and mining with coupled tensors
- 30 December 2017
- Published by Springer Nature
- p. 606-621
PowerCast: Mining and Forecasting Power Grid Sequences
- 30 December 2017
- Published by Springer Nature
- p. 606-621
PowerCast: Mining and Forecasting Power Grid Sequences
- 12 April 2016
- Published by Springer Nature
- p. 376-387
Matrices, Compression, Learning Curves: Formulation, and the GroupNteach Algorithms
- 12 April 2016
- Published by Springer Nature
- p. 376-387
Matrices, Compression, Learning Curves: Formulation, and the GroupNteach Algorithms