Where & When

Kista: 4 juni kl. 12:30-16:00
Göteborg: 5 juni kl. 08:30-12:00

Request more information

Alexander Majzoub
+ 46 40 59 22 03
am@nohau.se

På detta seminarium fördjupar vi oss ytterligare & visar hur AI-tekniken utvecklas. Nu skall mjukvaran appliceras vid utveckling av autonomi i mobila maskiner, där är maskininlärning i realtid nyckeln! En autonom maskin fungerar i en miljö där saker som inte alltid är logiska eller rationella händer. Hur kan maskinen få medel att klara sig och arbeta effektivt?

Vi tar upp några av utmaningarna med AI och ML samt visar på lösningar i form av färdiga programpaket för bl.a. Neurala nätverk. Praktisk tillämpning för objektigenkänning visar vi också och berättar hur du skapar.

Agenda:

– Individual and Incremental Machine Learning for Edge Devices. Rikard König, Ekkono Solutions
– Project Trillium, Unleashing Innovation for ML on Arm. Gilles Augustins, ARM
– Deep learning computer vision on a Cortex M7 with CMSIS-NN. Kalevi Ratschunas, Nohau


Individual and incremental machine learning for edge devices

Rikard KönigThis talk will focus on the possibilities and challenges that comes with moving AI/ML from the cloud out to edge devices. More specifically it will deal with the differences of centralized and decentralized learning and how the devices can learn individually over time using incremental learning. Finally, pros and cons of different incremental machine learning techniques for edge analytics will be discussed.

Speaker: Rikard König, PhD, is co-founder and CTO of Ekkono Solutions, and Senior Lecturer at the University of Borås, Sweden. At Ekkono solutions Rikard König leads development of a resource efficient machine learning library for IoT. In addition to the ability to run on very small devices Ekkono specializes on incremental and individual learning. At the University of Borås he is a member of CSL@BS, a research group of ten researchers specializing in machine learning and high-performance computing. Rikard König has published over 50 machine learning papers at international research conferences and journals.

Project Trillium, Unleashing Innovation for ML on Arm

Designed for unmatched versatility and scalability, Project Trillium is Arm’s heterogeneous ML platform. Project Trillium is advancing a new era of ultra-efficient ML inference at the edge, providing a range of performance options through a suite of products based on the world’s most innovative and advanced technologies.

Speaker: Gilles Augustins is managing sales of Arm Software and Development Tools in Europe, and is involved in shaping Arm Machine Learning tools strategy. He’s been with Arm for ten years, in many sales and business development positions, covering partners across Europe for CPUs, silicon libraries and software products.

Deep learning computer vision on a Cortex M7 with CMSIS-NN

Using software to parse the world’s visual content is as big of a revolution in computing as mobile was 10 years ago. Computer Vision is the process of using machines to understand and analyze imagery. The capability of executing machine learning at the edge will allow new types of services to be imagined which rely on low latency, privacy and being disconnected from the cloud. At this seminar I will introduce an efficient software based neural network acceleration implementation for computer vision.

Speaker: Kalevi Ratschunas’ career spans 20 years in the fields of smart phones, embedded and IoT. He has been part of great consumer products coming to life in the mobile and semiconductor industries. In his current role he works on machine learning projects and supports embedded SW development tools.


Var & När

> Kista: 4 juni kl. 12:30-16:00
Adress: Kista Entré konferens, Knarrarnäsgatan 7, Kista

> Göteborg: 5 juni kl. 08:30-12:00
Adress: Restaurang Ullevigatan, Ullevigatan 17-19, Göteborg

Pris: Fritt inträde – men anmälan krävs. Vi bjuder på fika i pausen.
Vi har begränsat antal platser så det är “först till kvarn…” som gäller.

Välkommen!

Alexander Majzoub
Tel: 040-59 22 03
am@nohau.se

Do you wish more information!

Contact us for offers, information or advice!