2nd Workshop on Embedded Machine Learning - WEML2018
Heidelberg University, Nov 8, 2018
INF306, SR14 (ground level)
Holger Fröning, ZITI, Heidelberg University, Germany
Franz Pernkopf, Graz University of Technology, Austria
Manfred Mücke, Materials Center Leoben, Leoben, Austria
In November 2018, our second Workshop on Embedded Machine (WEML) took place in Heidelberg, Germany. We received a nice attendance of more than 30 people interested in the broad topic of applying machine learning methods to embedded systems. Speakers and talks were as follows:
Schedule
Schedule
- 13:00 - 13:15 Holger Fröning/Heidelberg University: Introduction
- 13:15 - 13:50 Günther Schindler, Holger Fröning/Heidelberg University: The DeepChip Software Architecture
- 13:50 - 14:25 Wolfgang Roth, Franz Pernkopf/Graz University of Technology: Bayesian networks and PGMs
- 14:25 - 15:00 Christoph Gratl, Manfred Mücke/Materials Center Leoben: Investigating data-type propagation in Tensorflow/XLA-AOT/LLVM for ARM v7 architectures
- 15:00 - 15:45 Coffee break
- 15:45 - 16:20 Robert Peharz/Cambridge: Minimal Random Code Learning: Getting Bits Back from Compressed Model Parameters
- 16:20 - 16:55 Nick Fraser/Xilinx Ireland: Training Methods For Reduced Precision Neural Networks
- 16:55 - 17:30 Peter Zaspel/Basel University: Augmenting the explanatory power of predictions by uncertainty quantification
- 17:30 - 18:05 Jonas Große Sundrup/Heidelberg University: Towards embedded Machine Learning for motion classification in patient monitoring systems
- 18:05 - 18:20 Concluding discussion