Workshops inom maskininlärning
The seminar on brain analysis took place June 17th, 2022 at Luleå University of Technology.
The workshop on large-scale neuromorphic systems integration took place September 27th, 2022 at Luleå University of Technology.
Seminar on Brain analysis
Our first seminar will be on brain analysis techniques from the Neuroscience perspective and from the Machine Learning perspective using non invasive methods like Electroencephalography (EEG) and Functional magnetic resonance imaging (fMRI). Five outstanding speakers from Sweden, India and United Kingdom have been invited.
Program
09.00 – 09.20 Welcome – Opening remarks, Foteini Simistira Liwicki, LTU, SE
09.20 – 10.00 fMRI and its applications for visual decoding, Holly Wilson, UB, UK
10.00 – 10.30 Decoding Imagined Speech from the Brain: What is Possible, and Developments at the Frontier, Scott Wellington, UB, UK
10.30 – 10.45 Coffee Break
10.45 – 11.15 Cognitive workload estimation using EEG, Debashis Das Chakladhar, IIT Roorkee, IN
11.15 – 11.45 Time-frequency analysis methods and their application in developmental EEG data, Vibha Gupta, LTU, SE
11.45 – 13.00 Lunch
13.00 – 13.30 Pilot study on inner speech, Foteini Simistira Liwicki, LTU, SE
13.30 – 14.00 Information-theoretic and optimization approaches for EEG channel selection, Samar Agnihotri, IIT Mandi, IN
14.00 – 14.30 ECoG Hand Pose Estimation-Hackathon topic, Kanjar De, Rajkumar Saini, LTU, SE
14.30 – 15.00 A brief introduction to the neuroscience of consciousness, Johan Eriksson, UMU, SE
External invited speakers
Johan Eriksson
Associate Professor at Department of Integrative Medical Biology (IMB),
Umeå universitet, Sweden
A brief introduction to the neuroscience of consciousness
How the brain generates consciousness remains profoundly mysterious despite decades of dedicated research. I will give a brief overview of the current state of knowledge regarding the neural correlates of consciousness (NCC). The field still lacks a consensus on what the NCC is, due to both conceptual and methodological issues. The aim of the talk is to explicate these issues rather than to provide an answer to the problem of consciousness.
https://www.umu.se/en/staff/johan-eriksson/ Länk till annan webbplats, öppnas i nytt fönster.
Samar Agnihotri
Associate Professor at School of Computing and EE, IIT Mandi, India
A brief introduction to Information-theoretic and optimization approaches for EEG channel selection
Selecting the optimum set of electroencephalography (EEG) electrodes is an important problem to be addressed to minimize the cost, computational complexity, and data to be processed of an EEG deployment. This talk will provide a brief overview of information-theoretic and optimization-based approaches to address this problem and suggest some novel ideas to design better solutions for this problem based on these approaches.
http://faculty.iitmandi.ac.in/~samar/ Länk till annan webbplats, öppnas i nytt fönster.
Holly Wilson
Postgraduate Research Student, Department of Computer Science
University of Bath, United Kingtom
fMRI: and its applications for visual decoding
Talk description
Scott Wellington
Postgraduate Research Student, Department of Computer Science
University of Bath, United Kingtom
Decoding Imagined Speech from the Brain: What is Possible, and Developments at the Frontier
I will cover developments in the field: what has been achieved, what is being aimed for, etc. I'll cover my previous successes and show examples of what 'cutting-edge' results look like.
https://researchportal.bath.ac.uk/en/persons/scott-wellington Länk till annan webbplats, öppnas i nytt fönster.
Debashis Das Chakladar
PhD student, IIT Roorkee, India
Cognitive workload estimation using EEG
The workload can be defined as the physical and/or mental requirements associated with a task or combination of tasks. The workload can be divided into two types: Physical and cognitive workload. The cognitive workload is determined by mental stress and strain during the task. The level of cognitive workload (low, medium, hard) can be evaluated by subjective measures (NASA_TLX, SWAT), and physiological measures (EEG, fMRI, ECG, etc.). In this talk, I will highlight workload level estimation through different EEG-based tasks (mental arithmetic, n-back, etc.). The classification of workload level can be evaluated by several deep learning approaches.
Workshop on Large-Scale Neuromorphic Systems Integration
This workshop focuses on the challenges of integrating neuromorphic systems in large-scale digitized systems, for the purpose of improving the overall system-of-systems efficiency and artificial intelligence capabilities from core to edge.
Seminars by outstanding speakers from Sweden, United Kingdom, Germany, Switzerland, and the USA will cover different aspects of neuromorphic computing and integration.
The workshop is open to all scientists and students for physical attendance on campus at LTU. No sign-up is required.
Program
13:00 Welcome – Opening remarks, Fredrik Sandin, LTU, SE.
13:20 Experience of large-scale neuromorphic computing with SpiNNaker
Overview of SpiNNaker, Steve Furber, Manchester University, UK.
Software & Users, Andrew Rowley, Manchester University, UK.
Machine Learning, Michael Hopkins, Manchester University, UK.
14:00 Neuromorphic Computing Software,
Catherine Schuman, University of Tennessee, USA.
14:20 Scaling Neuromorphic Hardware on the Edge
Through Hardware-Algorithm Co-design,
Melika Payvand & Elisa Donati, Institute of Neuroinformatics, CH.
14:40 Concepts for Neuromorphic Systems Integration,
Mattias Nilsson, LTU, SE.
15:00 Coffee break
15:20 Efficient Neuromorphic Concepts for Automotive Applications,
Klaus Knobloch, Infineon, DE.
15:40 Neuromorphic Optimizer for Radio Access Network,
Ahsan Javed Awan, Ericsson Research, SE.
16:00 Process Orchestration with ColonyOS,
Johan Kristiansson, RISE, SE.
16:20 Round-table discussion,
moderator Olov Schelén, LTU, SE.
16:55 - 17:00 Summary and next steps, the organisers.
Organisers
Fredrik Sandin, Machine Learning, LTU.
Olov Schelén, Cyber-Physical Systems, LTU.
Ulf Bodin, Cyber-Physical Systems, LTU.
Anders Lindgren, Applied AI and IoT, RISE.
Sana Al-Azzawi, Machine Learning, LTU.
Uppdaterad:
Sidansvarig: Personal