Spiking neural networks: Applications to computing, algorithmics, and robotics

3rd HBP Curriculum Workshop Series - ICT for non-specialists

27-28 June 2019 | Technical University of Munich, Germany

 

 

FAQs & ALL YOU NEED TO KNOW

You have to expect the following expenses:

  • Registration fee (€ 250.00)
  • Travel
  • Accommodation
  • Breakfast & dinner

What we cover: 

  • Lunch & coffee breaks
  • Up to 5 fee waivers
     

Find further information on the application process and support opportunities on the .

In order to get to TUM Munich – Garching Campus you have many possibilities.

If you are coming from Munich Central Station:

  • Subway (U-Bahn)
    You can take the U-Bahn U1 line in the direction of Mangfallplatz or U2 line in the direction Messestadt Ost or U7 line in the direction of Neuperlach Zentrum until next stop, Sendlinger Tor. From here take the U-Bahn line U6 in the direction of Garching-Forschungszentrum, end station.
    Alternatively, every S-Bahn in the direction of Marienplatz or Ostbahnhof, and respectively only until Marienplatz (approx. 3 minutes). As of Marienplatz  U-Bahn U6 line until Garching Forschungszentrum (approx. 30 minutes).
     
  • S-Bahn
    You can take the S-Bahn S8 until you pass Ismaning. From here take Bus Nr. 230 until you stop at Garching Forschungszentrum. Alternatively: S1 line until Neufahrn, then with the Bus 690, which stops also at Garching Forschungszentrum.
     
  • Bus
    You can take the Buses Nr. 230 292 or 690 to Garching Forschungszentrum (Bus Nr. 230 only workdays in 40 minute intervals.) Buses wait up to 8 minutes for belated S-Bahn's.

If you are coming from Munich Airport:

  • You can take the S-Bahn S8 line in the direction of Munich Central Station until the third stop, Ismaning (approx. 13 minutes). From here Bus Nr. 230 until Garching Forschungszentrum
    Alternatively: S1 line until Neufahrn, then with the Bus 690.

We strongly recommend that you invest in travel insurance and inform yourself about treatment and reimbursement conditions before travelling to Munich, Germany. If you are a European citizen, you should bring your European Health Insurance Card, which is accepted in combination with an ID.
As health care and social security systems vary between EU countries, please check the details on unforeseen medical treatment abroad
Please note that you might have to contact your health insurer for authorisation before being treated at the hospital. 

Munich has continental climate. Summer is warm with average high temperatures of 23°C (73°F). June could be rainy too so pack a raincoat with you.

EDUROAM is available at Technical University of Munich.

  • Language:
    The official language in Germany is German. Foreign languages, particularly English, are widely understood and spoken.
     
  • Time zone:
    Munich is in the Central European Time Zone (CET = GMT / UCT + 1).
     
  • Electric current:
    Electricity is supplied at 230 volts (alternating current). Type F plugs (CEE 7/4, CEE 7/7) are used.

Please note that the information provided on this site has been obtained from several different sources and therefore the organisers cannot accept any responsibility for errors therein.

Spiking neural networks (SNN) are a special class of artificial neural networks (ANN), in which the information is transmitted by means of pulses (or spikes) rather than by firing rates. As SNNs have shown to be excellent control systems for biological organisms, they have the potential to produce good control systems for autonomous robots. This workshop aims to bring together neuroscientists with roboticists and computational researchers developing biologically-inspired learning algorithms for scientific and industrial applications.

This joint workshop between HBP and the EU-funded initiatives in robotics, TETRAMAX1 and TERRINet2, is a great opportunity for neuroscientists and HBP to reach out and engage with the robotics and research infrastructure communities. At the same time, it brings together students and early-career researchers in both areas fostering the exchange between them by focusing on recent developments in efficient neural coding and spiking neurons' computation.

In order to enrich the discussions of SNN and its applications to computing, algorithmics and robotics, we will have a hands-on session on spiking neural networks on the Neurorobotics Platform (HBP SP10) and a guided visit to TUM Competence Center.

 

Application deadline: 20 May 2019

 

APPLICATION

 

 

PROGRAMME

The scientific programme for download will be available soon.

Thursday 27 June 2019

Spiking neural networks as decision making mechanisms | 90 min
Yannick Morel (Technical University of Munich)
 

New methods for using spiking neural networks in robotics | 90 min
tbd
 

Spiking neural networks on the Neurorobotics Platform | 90 min
Fabrice Morin (Technical University of Munich)
 

Hands-on session: Spiking neural networks on the Neurorobotics Platform | 90 min
 

Poster session | 30 min

Friday 28 June 2019

Engineers inspire from nature: Modelling basal ganglia circuits with spiking neural networks | 120 min
Emeç Erçelik (Technical University of Munich), Neslihan Serap Şengör (Istanbul Technical University)
 

Spiking neural networks, human-robot interaction and cooperative manufacturing | 90 min
Esra Icer (Technical University of Munich)
 

Neural networks in robotics | 40 min
Michael Zechmair (Technical University of Munich)
 

Guided tour of TUM‘s Competence Center | 60 min
Esra Icer, Michael Zechmair (Technical University of Munich)

This programme may be subject to change.

 

CONFIRMED SPEAKERS

Ercelik picture Emeç Erçelik received his B.Sc. and M.Sc. degrees in Electronics Engineering from Istanbul Technical University, Istanbul, Turkey in 2013 and 2016, respectively. He started his Ph.D. in 2017 at the Technical University of Munich (TUM) in Germany and continues his Ph.D. at the Chair of Robotics, Artificial Intelligence and Real-time Systems at the same university. During his studies, he involved in different projects in neurorobotics as well as autonomous driving using spiking neural networks and artificial neural networks.

He worked as a Research Assistant between 2013 and 2016 at the Electronics Engineering Department of Istanbul Technical University. In 2017, he joined the Department of Electrical and Computer Engineering of TUM as a Research Assistant. Since April 2018, he has been working as a Research Assistant at the Chair of Robotics, Artificial Intelligence and Embedded Systems of Technical University of Munich.

 

Lecture title: Engineers inspire from nature: Modeling basal ganglia circuits with spiking neural networks

Brain is interesting; not only because understanding it will help us to be aware of ourselves but also will help us to build intelligent systems. Thus, engineers are now focusing on brain again; this time the inspiration is quite different than the artificial neural network era, which gave rise to machine learning. Now, engineers are no longer satisfied with mimicking the computational properties only, but they are also interested in building hardware inspired by brain. One way of doing this would be designing devices based on the biological specification of neurons as in Neurogrid. Another approach would be focusing on the computational properties and building computational systems as in SpiNNaker. In this talk, I will mention our humble attempt to understand the basal ganglia circuits and to make engineer wise use of our understanding. I will take you through our adventure where mathematics and computational methods are used together in building a spiking neural network model of basal ganglia circuit and how we used our results for getting Parkinson’s disease model and robotics applications.

Icer picture Esra Icer received her B.Sc. and M.Sc. degrees in Mechanical Engineering from Istanbul Technical University, Istanbul, Turkey in 2011 and 2013, respectively. She started her Ph.D. in 2014 at the Technische Universität München, Munich, Germany. During her studies, she involved in different projects in robotics.

She worked as a Research Assistant from 2011-2014 at the Mechanical Engineering Department of Istanbul Technical University. Between 2014 and 2017, she worked as an EU Researcher at the chair for Robotics and Embedded Systems, Department of Informatics, Technische Universität München. She was an Early Stage Researcher (ESR) involved in the EC FP7 Marie Curie Initial Training Network "Sustainable Manufacturing through Advanced Robotics Training in Europe (SMART-E)". Since February 2018, she has been working at the Technische Universität München as a Research Associate in ESMERA and TERRINet projects.

 

Lecture title: Spiking neural networks, human-robot interaction and cooperative manufacturing & Guided tour of TUM's Competence Center

By discussing the present and future perspectives of Spiking Neural Networks, we can better understand how SNN can be used as a cognitive model for the prediction of human intention to enable robots to effectively collaborate with their human co-workers in real-world manufacturing settings.

This will be complemented with a tour of the Technical University of Munich (TUM) Competence Center (CC), located at the Computer Science Department of TUM, Garching Hochbrück Campus. TUM CC consists of over 400 sqm of indoor space, which can be used for experimental testing of human-robot interaction and cooperative manufacturing, for the operation of mobile (including aerial) robots, and a Mobility Test Setup (MTS) for automotive industry.

The MTS in TUM CC provides repeatable testing conditions for tuning, learning and dynamic adaptation of traction control algorithms. Used in combination with the HBP-Neurorobotics Platform, which provides support for machine learning, including native support of Sony’s NNabla library and Google’s TensorFlow.

TUM CC has been used for many EU-funded projects such as HORSE, TERRINet, ESMERA, TETRAMAX and NeuroMotive.

Morel picture Yannick Morel has received a Master’s Degree in Electrical Engineering from the Institut Supérieur de l’ Electronique et du Numérique (ISEN), and a Master of Science in Ocean Engineering from Florida Atlantic University (FAU) in 2002. He was awarded a Ph.D. in Mechanical Engineering by Virginia Polytechnic Institute and State University (Virginia Tech, VT) in 2009.

He has worked in defense technology at the Institut de St-Louis (ISL, 2009-2010) and at Thales Underwater Systems (TUS, 2015-2016), and pursued research work on robotics at the Ecole Polytechnique Fédérale de Lausanne (EPFL, 2010-2014) and at the Commissariat à l’Energie Atomique (CEA, 2014-2015). He was Scientific Project Manager for the ECHORD++ project at the Technical University of Munich (TUM), and is now serving as Software Development Coordinator for the Human Brain Project.

Yannick Morel’s research interests include stability analysis of dynamical nonlinear systems, adaptive, nonlinear control theory with applications to robotics in general and motion control of unmanned vehicles in particular, probabilistic robotics, and multi-modal perception for unmanned vehicles.

 

Lecture title: Spiking neural networks as decision making mechanisms

Spiking Neural Networks use pulse coding mechanisms allowing to incorporate spatial temporal information. Better mimicking the nature of neural communication of biological neural networks, which gives the SSN the capacity to have richer dynamics and to exploit the temporal domain to encode or retrieve information in the exchanged spikes, and therefore the ability to learn and act in a dynamic environment. SNN enables the development of low-power computing architectures applicable to many fields, such as signal processing, classification, control, robotics and speech recognition.

Fabrice Morin received an engineering degree from the Institut National Polytechnique de Lorraine in 1999, and an MPhil in Bioengineering from the University of Strathclyde in 2001. Followed by a PhD degree from both the Ecole Nationale Superieure de Cachan (France) and the Japan Advanced Institute of Science and Technology (Japan) in 2004, focused in Biosensors based on cultured neural networks. He was then a JSPS post-doctoral fellow in the University of Tokyo, Japan (Fujita Laboratory) and a Marie Curie post-doctoral fellow at the IMS laboratory in Bordeaux, France.

Fabrice joined TUM’s Robotics and Embedded systems group in July 2017 as Scientific Coordinator for the HBP Director for Software Development, and as SP10 manager. His previous position was as a senior researcher in Neurotechnology in the Health division of Tecnalia, a private research center and nonprofit organization that is one the largest of its kind in Europe. There, his task was to design and implement R&D projects dealing with implantable active devices for interfacing with the peripheral nervous system, with special emphasis on the enteric nervous system. Between 2012 and 2015, Fabrice was manager of the business area of Biomaterials in Tecnalia.

 

Lecture title: Spiking neural networks on the Neurorobotics Platform

I will provide an overview of Spiking Neural Networks and brain-derived approaches in robotics and the NRP with its many features and benefits. Focusing on how the NRP enables connecting spiking neural networks to virtual and real robots. Though, either an easy to use web interface or on a local machine, allowing the user to build brains for bodies and bodies for brains.

Sengör picture Neslihan Serap Şengör interest is in nonlinear dynamical systems. She got her B.Sc., M.Sc. and PhD degrees from Istanbul Technical University (ITU), all in Electronics and Communication. Currently, she is holding full-professorship position in ITU and is working on computational neuroscience within Computational Neuroscience Group SIMMAG. She was visiting scientist at Circuit Theory Laboratory, Helsinki Technology University (now Aalto University) in 2000-2001. She returned to Helsinki Technology University during summer of 2006 as visiting scientist at Laboratory of Computational Engineering and worked on modelling cortex- basal ganglia- thalamus loop evaluating the reward. In the summer of 2015, she visited Lincoln University, UK as co-applicant of a project funded by Newton Funds for Research Collaboration, Royal Academy of Engineering and TÜBİTAK.

 

Lecture title: Engineers inspire from nature: Modeling basal ganglia circuits with spiking neural networks

Brain is interesting; not only because understanding it will help us to be aware of ourselves but also will help us to build intelligent systems. Thus, engineers are now focusing on brain again; this time the inspiration is quite different than the artificial neural network era, which gave rise to machine learning. Now, engineers are no longer satisfied with mimicking the computational properties only, but they are also interested in building hardware inspired by brain. One way of doing this would be designing devices based on the biological specification of neurons as in Neurogrid. Another approach would be focusing on the computational properties and building computational systems as in SpiNNaker. In this talk, I will mention our humble attempt to understand the basal ganglia circuits and to make engineer wise use of our understanding. I will take you through our adventure where mathematics and computational methods are used together in building a spiking neural network model of basal ganglia circuit and how we used our results for getting Parkinson’s disease model and robotics applications.

Zechmair picture Michael Zechmair received his B.Sc. and M.Sc. degrees in Electrical Engineering from the Technical University of Munich, Germany in 2014 and 2016, respectively. In 2018, he started his Ph.D. with a focus on modular gripper design.

 

Lecture title: Neural networks in robotics

This talk will give a brief overview of neural network use-cases in robotics. Automatization is taking over more and more areas which were previously only accomplishable by humans. Many of these tasks require complex operational sequences, dependent on external influences. In the last couple of years, neural networking has become a popular method for controlling robots. From manipulation to path planning, neural networks have made their mark on the area of task execution.

 

Lecture title: Guided tour of TUM's Competence Center

TUM CC consists of over 400 sqm of indoor space, which can be used for experimental testing of human-robot interaction and cooperative manufacturing, for the operation of mobile (including aerial) robots, and a Mobility Test Setup (MTS) for automotive industry.

The MTS in TUM CC provides repeatable testing conditions for tuning, learning and dynamic adaptation of traction control algorithms. Used in combination with the HBP-Neurorobotics Platform, which provides support for machine learning, including native support of Sony’s NNabla library and Google’s TensorFlow.

TUM CC has been used for many EU-funded projects such as HORSE, TERRINet, ESMERA, TETRAMAX and NeuroMotive.

SCIENTIFIC CHAIR

Alois Knoll | TUM

TUM logo

 

ORGANISERS

Esra Icer | TUM
Lisa-Marie Leichter | MUI
Yannick Morel | TUM
Fabrice Morin | TUM

 

ABOUT THE VENUE

TECHNICAL UNIVERSITY OF MUNICH
Munich/Garching
Germany
 

The university was founded in 1868 to provide the state of Bavaria with a center of learning dedicated to the natural sciences. The university played a vital role in Bavaria's transition from an agricultural to an industrial state – and accelerated the pace of technological advancement across Europe.
Now the Technical University of Munich is one of Europe's top universities. It is committed to excellence in research and teaching, interdisciplinary education and the active promotion of promising young scientists. The university also forges strong links with companies and scientific institutions across the world. TUM was one of the first universities in Germany to be named a University of Excellence. Moreover, TUM regularly ranks among the best European universities in international rankings.