The DeepChip Project - Deep Learning for Resource-Constrained Systems

DeepChip focusses on deep learning techniques for resource-constrained systems. Many processes require evaluation of complex numerical functions close to the machine or structure of interest, to avoid the effort of data transfer or to enable small reaction times. Although computing performance of embedded platforms is increasing, it is often significantly lower than the requirements of state-of-the-art algorithms. With the advent of Deep Neural Networks (DNN), the achievable classification performance has been pushed to new levels. The high cost of execution, however, renders them unusable to many real-world applications. A possible approach is the use of hybrid processors (ARM+FPGA or similar), but this raises the question on how to auto-generate optimized DNN classifier implementations. In the DeepChip project, we tackle this problem by optimizing deep models in terms of sparsity, asynchrony and reduced precision, and by extending machine learning languages with a hybrid back-end that is responsible for code generation, automated partitioning and integration.

The DeepChip Project was a FWF/DFG co-funded D-A-CH project, run by Graz University of Technology and Heidelberg University. While the first partner contributes expertise and experience from the machine learning area and applications, the second partner has a strong background on application-specific computing systems of various scale. Within the DeepChip project, the partners jointly pursue the objective of designing a productive and easy-to-use tool chain to design custom hardware for deep learning purposes, thereby contributing to bringing advanced machine learning techniques and principles to tiny embedded devices like mobile chips, Internet of Things and more.

Meantime, the term DeepChip serves as an umbrella for various activities in the context of embedded machine learning, with a particular focos on HW systems for machine learning, HW/SW codesign, and emerging technologies for learning-based applications.

For questions or comments, please contact: Holger Fröning, holger.froening (at) ziti.uni-heidelberg.de

Workshop mini-series on Embedded Machine Learning (WEML)

We are frequently hosting rather informal workshops that gather experts and interested people in machine learning, particulary deep learning, for embedded or other resource-constrained systems. More informations about recent incarnations can be found here:

• 2024 Workshop on Embedded Machine Learning (upcoming)

• 2023 Workshop on Embedded Machine Learning (past)

• 2019/20 Workshop on Embedded Machine Learning (past)

• 2018 Workshop on Embedded Machine Learning (past)

Workshop mini-series on IoT, Edge, and Mobile for Embedded Machine Learning (ITEM)

While WEML is a rather informal gathering with no proceedings or similar, ITEM is its academic counterpart, collocated ususally with ECML-PKDD as a premier European forum on ML. For more information about recent and upcoming editions, please visit www.item-workshop.org. The next edition is scheduled for September 2022.

Team

The DeepChip project is co-run by the 

• Computing Systems Group (http://www.ziti.uni-heidelberg.de/compsys, formerly Computer Engineering Group), Institute of Computer Engineering (http://www.ziti.uni-heidelberg.de) at  Heidelberg University, Germany (http://www.uni-heidelberg.de). 

• Signal Processing and Speech Communication Laboratory (https://www.spsc.tugraz.at) at Graz University of Technology (https://www.tugraz.at) 

Current people:

• Holger Fröning, Co-PI, Heidelberg University, Germany

• Franz Pernkopf, Co-PI, Graz University of Technology, Austria

• Bernhard Klein, PhD student, Heidelberg University, Germany

• Kevin Stehle, PhD student, Heidelberg University, Germany

• Hendrik Borras, PhD student, Heidelberg University, Germany

• Sophie Steger, PhD student, Graz University of Technology, Austria

• Daniel Barley, PhD student, Heidelberg University, Germany

Associated partners

• Manfred Mücke, Materials Center Leoben, Austria

Former people:

• Matthias Zöhrer, Graz University of Technology, Austria

• Günther Schindler, Heidelberg University, Germany

• Wolfgang Roth, Graz University of Technology, Austria

• Yannick Emonds, PhD student, Heidelberg University, Germany

• Dennis Rieber, PhD student, Bosch, Germany

Publications

Please find an up-to-date list at https://csg.ziti.uni-heidelberg.de/projects/2015-deepchip 

Sub-projects

Please check the Computing Systems Group's github site for various projects: https://github.com/UniHD-CEG

Sponsors

We gratefully acknowledge the sponsoring we receive from the DFG, FWF, FFG, Carl-Zeiss Foundation, and various others.