Call for Papers

Security für Deep Learning ( Vortrag )

Management von Safety und Security bei programmierbaren SoCs

Referent: Dr. James Hunt, aicas GmbH
Vortragsreihe: Machine Learning
Zeit: 06.12.17 16:45-17:25
Co-Referenten: Dr. Giulio Corradi is ISM (Industrial Scientific Medical) Sr. System Architect in Munich, Germany. He brings 25 year of experience of management, software engineering embedded systems and development of ASICs and FPGA in industrial, automation and medical systems specifically in the field of control and communication for the major corporations. Machine Learning, power system control, motor control, real-time communication, and functional safety have been his major focus. In the years 1997 – 2005 he managed several Research European Funded projects for train communication networking and wireless remote diagnostic systems. Between2000-2005 he headed the IEC61375 conformance test standard. In 2006 Giulio joined Xilinx in Munich (Germany) contributing to the Industrial Networking, Motor Control and Xilinx Functional Safety certification of Tools and compilers. In his spare time he practices swimming and playing piano.

Zielgruppe

Entwicklung

Themenbereiche

Implementierung, Sichere Software

Schwerpunkt

Technologie

Voraussetzungen

Grundlagenwissen

Kurzfassung

Machine Learning and artificial intelligence are big new trends impacting the future of embedded systems and are central to supporting intelligence and autonomous decision making. New paradigms form the core of Deep Learning algorithms. How can the use of these techniques in embedded control systems be eased while keeping up with the latest improvements for domain specific applications? There is a gap to be filled, the requirements of secure, dynamic update, reconfiguration, and monitoring will change the way these systems are deployed. A type-safe managed language platform can provide both the necessary safety and security needed for theses systems, but also ease of use for accessing and updating GPU and FPGA coded functions.This lecture will show how embedded engineers can easily learn the basics and implement solutions using Programmable SoC with new software-oriented design methodologies morphing the parallel hardware into a manageable and efficient development software process.

Gliederung

Presenting special requirements for machine learning and artificial intelligence in Hardware and Software with examples as well as paradigms for deep learning.
Description of needs for common consideration of security and safety for the HW and SW for machine learning systems
Demonstration of the Gap and presentation of solutions


Nutzen und Besonderheiten

In the lecture the whole complex correlation of performant HW- boards and ambitious SW-development for it will be spanned. With an example it will be demonstrated, how FPGA-development for machine-learning systems can be eased. It will cover the learning process, the design process, the advantages, and the challenges with real implementation examples demonstrating what these new paradigms bring to the embedded engineer.

Über den Referenten

Dr. James J. Hunt is a cofounder and CEO. He has a BS in Computer Science and Physics from Yale University, an MA in Computer Science from Boston University, and a Doctorate of Engineering in Computer Science from the University of Karlsruhe. He spent several years as a researcher at the Massachusetts Institute of Technology Lincoln Laboratory developing CAD software for restructurable wafer scale integrated circuits (RVLSI) and a parallel Lisp system for signal processing.