Project

Back to overview

Neuromorphic Attention

English title Neuromorphic Attention
Applicant Indiveri Giacomo
Number 121713
Funding scheme Project funding
Research institution Institut für Neuroinformatik Universität Zürich Irchel und ETH Zürich
Institution of higher education University of Zurich - ZH
Main discipline Electrical Engineering
Start/End 01.01.2009 - 31.12.2012
Approved amount 165'725.00
Show all

All Disciplines (2)

Discipline
Electrical Engineering
Neurophysiology and Brain Research

Keywords (15)

neuromorphic; attention; neural; network; VLSI; analog; multi-chip; active; vision; Selective Attention; Neural Network; Active Vision; Neuromorphic VLSI; Spiking Neuron; Synaptic Dynamics

Lay Summary (English)

Lead
Lay summary
Selective attention is an extremely effective strategy for building artificial systems, such as robots, active vision systems, or active action-perception systems, which require robust performance and real-time response properties.The aim of this project is to develop a real-time active selective-attention system by interfacing neuromorphic VLSI spiking neuron chips to neuromorphic sensors and robotic actuators. The neuromorphic approach implements the principles of computation used by the nervous systems in analog/digital VLSI technology, exploiting the device physics of transistors to reproduce the biophysics of neural cells. The neuromorphic chips model the properties of biological systems down to the single neuron and synapse dynamics and use a digital asynchronous communication infrastructure to transmit spikes across chips and to computers and robots.In addition to implementing artificial selective attention systems for practical applications, these systems represent additional basic research tools useful for validating neuro-biological models of selective attention and investigating the feasibility of a wide variety of neuroscience models and hypotheses.
Direct link to Lay Summary Last update: 21.02.2013

Responsible applicant and co-applicants

Employees

Publications

Publication
A Real-Time Event-Based Selective Attention System for Active Vision
Sonnleithner Daniel, Indiveri Giacomo (2012), A Real-Time Event-Based Selective Attention System for Active Vision, in Ulrich Rückert (ed.), Springer Berlin Heidelberg, Germany, 205-219.
A Neuromorphic Saliency-Map based Active Vision System
Sonnleithner Daniel, Indiveri Giacomo (2011), A Neuromorphic Saliency-Map based Active Vision System, in Conference on Information Sciences and Systems - CISS 2011, Washington, USAIEEE, New York.
Active vision driven by a neuromorphic selective attention system
Sonnleithner Daniel, Indiveri Giacomo (2011), Active vision driven by a neuromorphic selective attention system, in Proc. of International Symposium on Autonomous Minirobots for Research and Edutainment, AMiRE 2011, Bielefeld, GermanySpringer, Heidelberg.

Collaboration

Group / person Country
Types of collaboration
Italian Institute of Technology Italy (Europe)
- in-depth/constructive exchanges on approaches, methods or results

Scientific events

Active participation

Title Type of contribution Title of article or contribution Date Place Persons involved
International Symposium on Autonomous Minirobots for Research and Edutainment, AMiRE 2011 Talk given at a conference A Real-Time Event-Based Selective Attention System for Active Vision 23.05.2011 Bielefeld, Germany, Germany Sonnleithner Daniel;
The 2011 CapoCaccia Cognitive Neuromorphic Engineering Workshop Individual talk Mobile robots and selective attention chips 27.04.2011 Alghero, Italy, Italy Sonnleithner Daniel; Indiveri Giacomo;
Conference on Information Sciences and Systems Talk given at a conference A Neuromorphic Saliency-Map based Active Vision System 23.03.2011 Washington, USA, United States of America Sonnleithner Daniel;
The 2010 CapoCaccia Cognitive Neuromorphic Engineering Workshop Talk given at a conference Hands-on projects with Neuromorphic Attention chips 25.04.2010 Alghero, Italy Sonnleithner Daniel; Indiveri Giacomo;


Communication with the public

Communication Title Media Place Year
Talks/events/exhibitions BrainFair 2012 "Gehirn und Technologie" German-speaking Switzerland 2012

Associated projects

Number Title Start Funding scheme
119973 Real-time sound recognition using neuromorphic VLSI 01.04.2009 Project funding

Abstract

Biological organisms take complex decisions continuously and effortlessly. These operations allow them to quickly determine the motor actions to take in response to combinations of external stimuli and internal states, or to pay attention to subsets of sensory inputs suppressing non salient ones, or to plan complex action sequences serially choosing elementary behaviors among different alternatives. The mechanism that enables biological systems to perform so efficiently in this respect is that of "selective attention". This extremely powerful strategy can be used by artificial systems, such as robots, or embedded systems, for optimizing performance and achieving results that could not be accomplished otherwise. Indeed, several attempts have been made to incorporate attentional strategies into artificial computational systems. However, given the complexity of the processing involved, achieving robust performance in real-time has required bulky, expensive and power-hungry computing systems, that are not ideal for robotic applications or potential commercial solutions. Furthermore, as compromises have often been made between the level of detail used in modeling the biological features of selective attention, and the complexity and processing power devoted to the simulation tools, these attempts have also not been ideal for basic research investigations. We propose to develop a full-fledged audio/visual selective attention mechanism using custom "neuromorphic" VLSI devices, able to operate in real-time, with low power, and that can be packaged into an extremely compact and light-weight system. The neuromorphic approach aims to implement the principles of computation used by the nervous systems in analog/digital VLSI technology, exploiting the device physics of transistors to reproduce the biophysics of neural cells. The neuromorphic devices that we will use model in great detail the properties of biological selective attention systems down to the single neuron and single synapse dynamics. Therefore these systems will also be additional valuable tools for validating neuro-biological models of selective attention and investigating the feasibility of different neuroscience models and hypotheses.
-