********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +**********************************************************************
Zeit: Dienstag, 15. Dezember 2020, 10:00 Uhr Zoom: https://rwth.zoom.us/j/99233095930?pwd=dHhTV253V1ZYUzRtSkk1L3A1REZVUT09
Meeting-ID: 992 3309 5930 Kenncode: 626162
Referent: Philipp Weidel, Dipl. Inform.
Thema: Learning and decision making in closed loop simulations of plastic spiking neural networks
Abstract:
To understand how animals and humans learn, form memories and make decisions is a highly relevant goal both for neuroscience and for fields that take some inspiration from neuroscience, such as machine learning and artificial intelligence. Many models of learning and decision making were developed in the fields of machine learning, artificial intelligence, and computational neuroscience. Although these models aim to describe similar mechanisms, they do not all pursue the same goal. These models can be differentiated between models aiming to reach optimal performance on a specific task (or set of tasks) and models trying to explain how animals and humans learn. Some models of the first class use biologically inspired methods (such as deep learning) but are usually not biologically realistic and are therefore not well suited to explain the function of the brain. Models in the second class focus on being biologically plausible to explain how the brain works, but often demonstrate their capability on too simplistic tasks and yield low performance on well-known tasks from machine learning. This work aims to close the gap between these two types of models.
In the first part of this talk, tools are described that allow the combination of biologically plausible neural network models together with powerful toolkits known from machine learning and robotics. To this end, MUSIC, the middleware for spiking neural network simulators such as NEST and NEURON is interfaced with ROS, a middleware for robotic hardware and simulators such as Gazebo. This toolchain is extended with interfaces to reinforcement learning toolkits such as the OpenAI Gym.
The second part addresses the question of how the brain can represent its environment in the neural substrate of the cortex and how a realistic model of reinforcement learning can make use of these representations. To this end, a spiking neural network model of unsupervised learning is presented which is able to learn its input projections such that it can detect and represent repeating patterns. By using an actor-critic reinforcement learning architecture driven by a realistic dopamine modulated plasticity rule the model can make use of the representations and learn a range tasks.
Es laden ein: die Dozentinnen und Dozenten der Informatik
-- Prof. Dr. Abigail Morrison
IAS-6 / INM-6 / SimLab Neuroscience Jülich Research Center
&
Computer Science 3 - Software Engineering RWTH Aachen
http://www.fz-juelich.de/inm/inm-6 http://www.fz-juelich.de/ias/jsc/slns http://www.se-rwth.de
Office: +49 2461 61-9805 Fax # : +49 2461 61-9460
------------------------------------------------------------------------------------------------ ------------------------------------------------------------------------------------------------ Forschungszentrum Juelich GmbH 52425 Juelich Sitz der Gesellschaft: Juelich Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498 Vorsitzender des Aufsichtsrats: MinDir Volker Rieke Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender), Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt ------------------------------------------------------------------------------------------------ ------------------------------------------------------------------------------------------------
Zeige Antworten nach Diskussionsstrang
********************************************************************** * * * Einladung * * * * Informatik-Oberseminar * * * +**********************************************************************
Zeit: Thursday, 11. March 2021, 10:00
Zoom: https://rwth.zoom.us/j/92520089251?pwd=c1RwbUF4dzF4WWdDVWlBMDlHOFlZdz09
Meeting-ID: 925 2008 9251 Kenncode: 269134
Referent: Sandra Diaz Pier, M.Sc. Eng.
Thema: Structural plasticity as a connectivity generation and optimization algorithm in neural networks
Abstrakt:
In many fields of science, models are based on sets of differential equations which need to be fit against experimental data. In order to do this, parameter spaces are searched to find specific values which make these models useful to answer relevant scientific questions. In computational neuroscience, models of spiking networks of neurons play an important role in understanding how the brain encodes information and achieves high level cognitive functions. These models are not only of interest for neuroscience but also to many other related fields including artificial intelligence, robotics and control. However, these models are very underconstrained, degenerate and show chaotic dynamics which makes it challenging to find suitable and robust solutions.
In this presentation I propose structural plasticity as an optimization algorithm inspired by neurobiology able to generate, modify and tune connectivity parameters for neural network models. Structural plasticity refers to the ability of neurons to change their structure by creating and deleting connections with other neurons in a network in order to preserve specific metabolic levels. First, I will introduce the characteristics of structural plasticity as an optimization algorithm together with details about its implementation in NEST, a well-known neural network simulator within the computational neuroscience community. This implementation can efficiently leverage computational resources and is applicable to large scale neural networks. I will also briefly present a tool which I have co-developed in order to visualize, analyze and interact with simulations using structural plasticity. I will focus on how this tool can be used to better understand the relationships between structure and function emerging in neural networks.
The rules under which structural plasticity operates in the brain have been tuned through centuries of natural evolutionary optimization. In the second part of my talk I present how meta-optimization can be used to artificially explore the general rules which make structural plasticity able to work with a variety of network configurations and reach different functional regimes at each portion of the network. Finally, I will present different applications for structural plasticity in and outside neuroscience as well as future research directions.
Es laden ein: die Dozentinnen und Dozenten der Informatik
-- Prof. Dr. Abigail Morrison
IAS-6 / INM-6 / SimLab Neuroscience Jülich Research Center
&
Computer Science 3 - Software Engineering RWTH Aachen
http://www.fz-juelich.de/inm/inm-6 http://www.fz-juelich.de/ias/jsc/slns http://www.se-rwth.de
Office: +49 2461 61-9805 Fax # : +49 2461 61-9460 Pronouns: she/her
------------------------------------------------------------------------------------------------ ------------------------------------------------------------------------------------------------ Forschungszentrum Juelich GmbH 52425 Juelich Sitz der Gesellschaft: Juelich Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498 Vorsitzender des Aufsichtsrats: MinDir Volker Rieke Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender), Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt ------------------------------------------------------------------------------------------------ ------------------------------------------------------------------------------------------------
informatik-vortraege@lists.rwth-aachen.de
-
Abigail Morrison