TA1b: Architectures for Machine Learning
Tue, 31 Oct, 10:15 - 11:55 PT (UTC -8)
Location: Evergreen
Session Type: Lecture
Session Chair: Milos Ercegovac, University of California, Los Angeles
Track: Architectures and Implementation
Tue, 31 Oct, 10:15 - 10:40 PT (UTC -8)

TA1b.1: A Hardware-Oriented QAM Demodulation Method Driven by AW-SOM Machine Learning
Lorenzo Canese, Gian Carlo Cardarilli, Luca Di Nunzio, Rocco Fazzolari, Marco Re, Sergio Spanò, University of Rome, Italy
Tue, 31 Oct, 10:40 - 11:05 PT (UTC -8)

TA1b.2: Synaptic Turnover Promotes Efficient Learning in Bio-Realistic Spiking Neural Networks
Nikos Malakasis, Spyridon Chavlis, Panayiota Poirazi, Foundation for Research and Technology-Hellas (FORTH), Greece
Tue, 31 Oct, 11:05 - 11:30 PT (UTC -8)

TA1b.3: An Efficient Dot-Product Unit Based on Online Arithmetic for Variable Precision Applications
Saeid Gorgin, MohammadH. Golamrezaei, Jeong-A Lee, Chosun University, Republic of Korea; Miloˇs D. Ercegovac, University of California, Los Angeles, United States
Tue, 31 Oct, 11:30 - 11:55 PT (UTC -8)

TA1b.4: MSDF-SVM: Advantage of Most Significant Digit First Arithmetic for SVM Realization
Saeid Gorgin, Mohammadreza Najafi, Mohammad H. Golamrezaei, Jeong-A Lee, Chosun University, Republic of Korea; Miloˇs D. Ercegovac, University of California, Los Angeles, United States