Computers with Brains? A neuroscience perspective
description
Transcript of Computers with Brains? A neuroscience perspective
1
Computers with Brains? A neuroscience perspective
Khurshid Ahmad, Professor of Computer Science,
Department of Computer ScienceTrinity College,
Dublin-2, IRELANDOctober 29th , 2009.
https://www.cs.tcd.ie/Khurshid.Ahmad/Teaching/ComputersBrains.pdf
2
Brain – The Processor!
http://www.cs.duke.edu/brd/Teaching/Previous/AI/pix/noteasy1.gif
3
Brain – The Processor!Neural computing systems are trained on the principle that if a network can compute then it will learn to compute. Multi-net neural computing systems are trained on the principle that if two or more networks learn to compute simultaneously or sequentially , then the multi-net will learn to compute.
Our project is to build a neural computing system comprising networks that can not only process unisensory input and learn to process but that the interaction between networks produces multisensory interaction, integration, enhancement/suppression, and information fusion.
Tim Shallice (2006). From lesions to cognitive theory. Nature Neuroscience Vol 6, pp 215 (Book Review: Mark D’Esposito (2002). Neurological Foundations of Cognitive Neuroscience
4
What humans do?
5
What animals do?
Dendritic computation. The task of a brainstem auditory neuron performing coincidence detection in the sound localization system of birds is to respond only if the inputs arriving from both ears coincide in a precise manner (10–100 μs), while avoiding a response when the input comes from only one ear.
London, Michael and Michael Häusser (2005). Dendritic Computation. Annual Review of Neuroscience. Vol. 28, pp 503–32
6
What computer scientists do?
The study of the behaviour of neurons, either as 'single' neurons or as cluster of neurons controlling aspects of perception, cognition or motor behaviour, in animal nervous systems is currently being used to build information systems that are capable of autonomous and intelligent behaviour.
7
What animals do?
Neurons are known to be involved in much more sophisticated computations, such as face recognition [..]. An algorithm to solve a face recognition task is one of the holy grails of computer science. At present, we do not know precisely how single neurons are involved in this computation. An essential first step is feature extraction from the image, which clearly involves a lot of network pre-processing before features are fed into the individual cortical neuron. The flowchart implements a three-layer model of dendritic processing […] to integrate the input. The way such a flowchart is mapped onto the real geometry of a cortical pyramidal neuron remains unknown.
London, Michael and Michael Häusser (2005). Dendritic Computation. Annual Review of Neuroscience. Vol. 28, pp 503–32
8
What animals do?
London, Michael and Michael Häusser (2005). Dendritic Computation. Annual Review of Neuroscience. Vol. 28, pp 503–32
Neurons, and indeed networks of neurons perform highly specialised tasks. The dendrites bring the input in, the soma processes the input and then the axon outputs. However, it appears that the dendrites also have processing power: it is the equivalent of the wires that connects your computer to its printer and the network hub performing computations – helping the computer to perform computations!!!
9
What humans think about what computers will do?
http://www.longbets.org/1
10
What humans do? They have an intricate
brain
11http://en.wikipedia.org/wiki/Neural_network#Neural_networks_and_neuroscience
Neural Networks &Neurosciences
Observed Biological Processes (Data)
Biologically PlausibleMechanisms for Neural Processing & Learning
(Biological Neural Network Models)
Theory(Statistical Learning Theory &
Information Theory)
Neural Nets and Neurosciences
12
Real Neuroscience
Cognitive neuroscience has many intellectual roots.
The experimental side includes the very different methods of systems neuroscience, human experimental psychology and, functional imaging.
The theoretical side has contrasting approaches from neural networks or connectionism, symbolic artificial intelligence, theoretical linguistics and information-processing psychology.
Tim Shallice (2006). From lesions to cognitive theory. Nature Neuroscience Vol 6, pp 215 (Book Review: Mark D’Esposito (2002). Neurological Foundations of Cognitive Neuroscience
13
Real Neuroscience
London, Michael and Michael Häusser (2005). Dendritic Computation. Annual Review of Neuroscience. Vol. 28, pp 503–32
Brains compute. This means that they process information, creating abstract representations of physical entities and performing operations on this information in order to execute tasks. One of the main goals of computational neuroscience is to describe these transformations as a sequence of simple elementary steps organized in an algorithmic way. The mechanistic substrate for these computations has long been debated. Traditionally, relatively simple computational properties have been attributed to the individual neuron, with the complex computations that are the hallmark of brains being performed by the network of these simple elements.
14
Real Neuroscience
I compute, therefore I am (intelligent):
I can converse in natural languages;I can analyse images, pictures (comprising images), and scenes (comprising pictures);I can reason, with facts available to me, to infer new facts and contradict what I had known to be true;I can plan (ahead);I can use symbols and analogies to represent what I know;I can learn on my own, through instruction and/or experimentation;I can compute trajectories of objects on the earth, in water and in the air;I have a sense of where I am physically (prio-perception)I can deal with instructions, commands, requests, pleas;I can ‘repair’ myself;I can understand the mood/sentiment/affect of people and groupsI can debate the meaning(s) of life;
15
Real Neuroscience
I cannot compute, therefore I am (not intelligent?):
I cannot add/subtract/multiply/divide with consistent accuracy;I forget some of the patterns I had once memorised;I confuse facts;I cannot recall immediately what I know;I cannot solve complex equations;I am influenced by my environment when I make decisions, ask questions, pass comments;I will (eventually) loose my faculties and then die!!
16
Real Neuroscience
I can and annot not compute because of my nervous system
I use language, can see things, represent knowledge, learn, plan, reason …. Because of my nervous system?
I cannot do arithmetic consistently, cannot recall and/or forget…. Because of my nervous system?
17
DEFINITIONS: Artificial Neural Networks
Artificial Neural Networks (ANN) are computational systems, either hardware or software, which mimic animate neural systems comprising biological (real) neurons. An ANN is architecturally similar to a biological system in that the ANN also uses a number of simple, interconnected artificial neurons.
18
DEFINITIONS: Artificial Neural Networks
Artificial neural networks emulate threshold behaviour, simulate co-operative phenomenon by a network of 'simple' switches and are used in a variety of applications, like banking, currency trading, robotics, and experimental and animal psychology studies.
These information systems, neural networks or neuro-computing systems as they are popularly known, can be simulated by solving first-order difference or differential equations.
19
The real neurons are different!
Real neurons co-operate, compete and inhinbit each other. In multi-modal information processing, convergence of modalities is critical.
From Alex Meredith, Virginia Commonwealth University, Virginia, USA
MultisensoryEnhancement
Cross-modalSuppression
Cross-modalFacilitation
Cross-modalFacilitationInhibition-Dependent
20
Computation and its neural basis
Much of modern computing relies on the discrete serial processing of uni-modal data
Much of the computing in the brain is on sporadic, multi-modal data streams
21
Computation and its neural basis
Analysis of neuroscience experiments is carried out with simple models without the capability of learning
Neural computing emphasises the learning aspects of behaviour
22
Computation and its neural basis
Decision making invariably involves fusion of multi-modal data streams in the brain involving emotion and reasoning
Learning to make decisions in noisy environments is something very human; key areas are image annotation, sentiment analysis
23
Computation and its neural basis
Network Epicentre 1 Epicentre 2
Spatial awareness Posterior Parietal Cortex Frontal eye fields
Language Wernicke’s Area Broca’s area
Explicit memory/emotion Hippocampal-entorhinal complex The Amygdla
Face-Object recognition Mid-temporal cortex Temporo-polar cortex
Working memory-executive function Lateral pre-frontal cortex Posterior Parietal Cortex (?)
24
What computers can do? Artificial Neural Networks In a restricted sense artificial neurons are simple emulations of biological neurons: the artificial neuron can, in principle, receive its input from all other artificial neurons in the ANN; simple operations are performed on the input data; and, the recipient neuron can, in principle, pass its output onto all other neurons.
Intelligent behaviour can be simulated through computation in massively parallel networks of simple processors that store all their long-term knowledge in the connection strengths.
25
What computers can do? Artificial Neural Networks
According to Igor Aleksander, Neural Computing is the study of cellular networks that have a natural propensity for storing experiential knowledge.
Neural Computing Systems bear a resemblance to the brain in the sense that knowledge is acquired through training rather than programming and is retained due to changes in node functions.
Functionally, the knowledge takes the form of stable states or cycles of states in the operation of the net. A central property of such states is to recall these states or cycles in response to the presentation of cues.
26
DEFINITIONS: Artificial Neural Networks
27
What can computers do?
http://search.eb.com.elib.tcd.ie/eb/art-68188/Moores-law-In-1965-Gordon-E-Moore-observed-that-the
Computers help you in visualising the un-built environment
28
What can computers do?
http://search.eb.com.elib.tcd.ie/eb/art-68188/Moores-law-In-1965-Gordon-E-Moore-observed-that-the
Computers help you in visualising the un-built environment
29
What can computers do?
http://search.eb.com.elib.tcd.ie/eb/art-68188/Moores-law-In-1965-Gordon-E-Moore-observed-that-the
Computers help you in visualising the un-built environment
30
What can computers do?
http://www.nytimes.com/2009/07/26/science/26robot.html
Robots, containing sophisticated computers, can perform a variety of tasks in an automated factory, in bomb disposal and other applications. Robots need to be looked after – be energized, helped to navigate, tasks to be assigned, but, but ……….
31
What can computers do?
http://www.nytimes.com/2009/07/26/science/26robot.html & http://www.imdb.com/media/rm969447936/tt0062622
Robots, containing sophisticated computers, can perform a variety of tasks in an automated factory, in bomb disposal and other applications. Robots need to be looked after – be energized, helped to navigate, tasks to be assigned, but, but ……….
32
What can computers do?
33
What can computers do?
34
What can computers do?
http://www.ibm.com/ibm/ideasfromibm/us/roadrunner/20080609/index.shtml
Giga
Tera
RoadRunner = 100,000 Laptops (running simultaneously and exchanging information pico second by pico second)
35
Computers dealing with problems in neuroscience
http://affect.media.mit.edu/pdfs/07.Teeters-sm.pdf
Dealing with idiosyncratic behaviour: Letting the autistic child/adult have feedback on self
36
Computers dealing with problems in neuroscience
http://affect.media.mit.edu/pdfs/07.Teeters-sm.pdf
Dealing with idiosyncratic behaviour
37
How do computers do what computers do?
http://search.eb.com.elib.tcd.ie/eb/art-68188/Moores-law-In-1965-Gordon-E-Moore-observed-that-the
The number of chips on the same area has doubled every 18-24 months; and has increased exponentially. However, the R&D costs and manufacturing costs for building ultra-small, high-precision circuitry and controls has had an impact on the prices
38
The ever growing computer systems (1997)
http://www.transhumanist.com/volume1/moravec.htm
39
The ever growing computer systems (1997)
http://www.transhumanist.com/volume1/moravec.htm
40
The ever growing computer systems (1997)
http://www.transhumanist.com/volume1/moravec.htm
41
The ever growing computer systems
http://en.wikipedia.org/wiki/File:Cmos-chip_structure_in_2000s_(en).svghttp://www.transhumanist.com/volume1/moravec.htm & http://www.loc.gov/webarchiving/faq.html
The industry took 35 years to reach 1GB (in 1991), 14 years more to reach 500GB (in 2005), and just two more years to reach 1TB (2007)
Giga
Tera
Library of Congress Minerva Web Presence
42
The ever growing computer systems
http://search.eb.com.elib.tcd.ie/eb/art-68188/Moores-law-In-1965-Gordon-E-Moore-observed-that-the
The cost of computation is falling dramatically – an exponential decay in what we can get by spending $1000 (calculations per second):
In 1940: 0.01In 1950: 1In 1960: 100In 1970: 500-1000In 1980: 10,000In 1990: 100,000In 2000: 1,000,000
43
The ever growing computer systems
http://search.eb.com.elib.tcd.ie/eb/art-68188/Moores-law-In-1965-Gordon-E-Moore-observed-that-the
The cost of data storage is falling dramatically – an exponential decay in what we can get by spending the same amount of money:In 1980: 0.001 GBIn 1985: 0.01In 1990: 0.1In 1995: 1In 2000: 10In 2005: 100In 2010: 1000
44
The ever growing computer systems:
Supercomputers of today
300 to 1400 Trillion Floating Point Operations per Second
http://www.transhumanist.com/volume1/moravec.htm
45
The intelligent computer:Turing Test
http://en.wikipedia.org/wiki/File:Cmos-chip_structure_in_2000s_(en).svg
46
What humans think about what computers will do?
http://www.longbets.org/1
47
What can computers do?
48
What can computers do?
http://crca.ucsd.edu/~hcohen/cohenpdf/furtherexploits.pdf
Aaron, an expert system that can ‘paint’ has had ITS exhibition in major galleries Here is Aaron’s Liberty & Freinds based on the Statute of Liberty
49
What can computers do?
http://crca.ucsd.edu/~hcohen/cohenpdf/furtherexploits.pdf
Aaron, an expert system that can ‘paint’ has had ITS exhibition in major galleries Here is Aaron’s Theo
50
What can computers do?
http://crca.ucsd.edu/~hcohen/cohenpdf/furtherexploits.pdf
51
What can computers do?
http://crca.ucsd.edu/~hcohen/cohenpdf/furtherexploits.pdf
52
What computers cannot do?
Hans Moravec (1998). When will computer hardware match the human brain? Journal of Evolution and Technology. 1998. Vol. 1 (at http://www.transhumanist.com/volume1/moravec.htm)
The Vision Problem 1967-1997Thirty years of computer vision reveals that
1 MIPS can extract simple features from real-time imagery--tracking a white line or a white spot on a mottled background.
10 MIPS can follow complex gray-scale patches--as smart bombs, cruise missiles and early self-driving vans attest.
100 MIPS can follow moderately unpredictable features like roads--as recent long NAVLAB trips demonstrate.
1,000 MIPS will be adequate for coarse-grained three-dimensional spatial awareness--.
10,000 MIPS can find three-dimensional objects in clutter--
53
What computers cannot do?
http://www.electronicspecifier.com/Industry-News/New-SH7724-processors-add-HD-video-playback-and-recording-support-to-Renesas-Technologys-popular-SH772x-series-of-low-power-multimedia-processors.asp
The Vision Problem – The story continues (2009)
54
What computers cannot do?
http://www.electronicspecifier.com/Industry-News/New-SH7724-processors-add-HD-video-playback-and-recording-support-to-Renesas-Technologys-popular-SH772x-series-of-low-power-multimedia-processors.asp
The Vision Problem – The story continues (2009)Processors add HD video playback and recording support to Renesas Technology's popular SH772x series of low power multimedia processors
News Release from: Renesas Technology Europe Ltd27/05/2009
Renesas has announced the release of the SH7724, the third product in the SH772x series of low power application processors designed for multimedia applications such as audio and video for portable and industrial devices.
When operating at 500 MHz, general processing performance is 900 million instructions per second (MIPS) and FPU processing performance is 3.5 giga [billion] floating-point operations per second (GFLOPS).
55
What a smart computer does:
Represents knowledge!A Hierarchical Network•
salmonlays eggs; swims upstream,
is pink, is edible
ostrichruns fast, cannot fly,
is tall
canarycan sing, is yellow
birdcan fly, has wings,
has feathers
fishcan swim, has fins, has gills
animalcan breathe, can eat,
has skin
is-a
is-a
is-a
is-a
is-a
56
What a smart computer does:Represents knowledge and reasons with
it!
DOG SUGAR MAPLE
BREAD MOULD
INTESTINAL BACTERIUM
POND ALGAE
KINGDOM Animalia (animals)
Plantae (plants)
Fungi (fungi)
Prokaryoke (bacteria)
Protoctista (algae, protozoa, slim moulds)
PHYLUM Chordata Magnoliphyk Zygomycota Omnibacteria Chlorophyta
CLASS Mammalia Rosidae Zygomycetes Enterobacteria Evanjugatae
FAMILY Canidae Aceraceae Mucoraceae (E. coli does not have a family classification)
Zygnematale
GENUS Canis Acer Rhizopis Escherichia Zygnemataceae
SPECIES C. familaris A. saccharum R. stolonifer E. coli S. crassa
ORDER Carnivora Sapindale Macorales Eubacteriales Zygnematales
Modern taxonomy recognises five kingdoms, into which the five million species of the world are organised:
57
What humans do? Create new knowledge and keep contradictions
in
Modern taxonomy recognises five kingdoms, into which the five million species of the world are organised and then goes onto incorporate, in this fixed species world,
EVOLUTION Cladistics, or phylogenetic systematics, for me and you! Willi Henning
proposed a method for implementing Darwin’s concepts of ancestors and
descendants in a taxonomic description.
58
What humans do? Create new knowledge and keep contradictions
in
Modern taxonomy recognises five kingdoms, into which the five million species of the world are organised and then goes onto incorporate, in this fixed species world, EVOLUTION Cladistics, or
phylogenetic systematics, for me and you! Willi Henning proposed a method for implementing Darwin’s concepts of ancestors and descendants in a taxonomic description.
LAMPREY SHARK SALMON LIZARD D A B C
x
y
z t0
t1
t2
Lizard (LZ) & salmon (SN) are more closely related to each other is to shark (SK) LZ & SN share a common ancestor
59
What humans think about what computers do?
Minerva is a robot which was evaluated in action by people outside a laboratory
60
What humans think about what computers do?
Minerva is a robot which was evaluated in action by people outside a laboratory
61
Language can be viewed as 'a communicative process based on knowledge. Generally when humans use language, the producer and comprehender are processing information, making use of their knowledge of the language and of the topics of conversation. Language is a process of communication between intelligent active processors, in which both the producer and the comprehender(s) perform complex cognitive tasks
Winograd, Terry. (1983). Language as a Cognitive Process. Wokingham: Addison-Wesley Inc.,
What humans do? Talk, listen, read and write
62
Producer
Current Goals
Cognitive Processing
Knowledge Base
Knowledge of the language
Knowledge of the
situation
Knowledge of the world
Comprehender
Understood Meaning
Cognitive Processing
Knowledge Base
Knowledge of the language
Knowledge of the
situation
Knowledge of the world
Medium
Speech
or
Writing
What humans do? Talk, listen, read and write
63
What humans do? Talk, listen, read and write
Phonlogical Rules
Morphological Rules
Dictionary (items)
Grammar Rules
Dictionary Definitions
Semantic Rules
Deductive Rules
Inferential Rules
Phonological
Morphological
Lexical
Syntactic (Parsing)
Semantic
Reasoning
Sounds
Phonemes
Morphemes
Words
Syntactic Structures
Representation Structures
Stored Knowledge
Processes Assigned Structures
64
DEFINITIONS: Artificial Neural Networks In a restricted sense artificial neurons are simple emulations of biological neurons: the artificial neuron can, in principle, receive its input from all other artificial neurons in the ANN; simple operations are performed on the input data; and, the recipient neuron can, in principle, pass its output onto all other neurons.
Intelligent behaviour can be simulated through computation in massively parallel networks of simple processors that store all their long-term knowledge in the connection strengths.
65
DEFINITIONS:Neurons & Appendages
AXONCELL BODY
NUCLEUS
DENDRITES
A neuron is a cell with appendages; every cell has a nucleus and the one set of appendages brings in inputs – the dendrites – and another set helps to output signals generated by the cell
66
DEFINITIONS:Neurons & Appendages
AXONCELL BODY
NUCLEUS
DENDRITES
A neuron is a cell with appendages; every cell has a nucleus and the one set of appendages brings in inputs – the dendrites – and another set helps to output signals generated by the cell
The Real McCoy
67
DEFINITIONS:Neurons & Appendages
The human brain is mainly composed of neurons: specialised cells that exist to transfer information rapidly from one part of an animal's body to another.
This communication is achieved by the transmission (and reception) of electrical impulses (and chemicals) from neurons and other cells of the animal. Like other cells, neurons have a cell body that contains a nucleus enshrouded in a membrane which has double-layered ultrastructure with numerous pores.
Neurons have a variety of appendages, referred to as 'cytoplasmic processes known as neurites which end in close apposition to other cells. In higher animals, neurites are of two varieties: Axons are processes of generally of uniform diameter and conduct impulses away from the cell body; dendrites are short-branched processes and are used to conduct impulses towards the cell body.
The ends of the neurites, i.e. axons and dendrites are called synaptic terminals, and the cell-to-cell contacts they make are known as synapses.
SOURCE:http://en.wikipedia.org/wiki/Neurons
Nucleus
Dendrite
Soma
AxonTerminals
68
DEFINITIONS: The fan-ins and fan-outs
–
– +
summation
10 fan-in4
1 - 100 meters per sec.
Asynchronousfiring rate, c. 200 per sec.
10 fan-out4
1010 neurons with 104 connections and an average of 10 spikes per second = 1015 adds/sec. This is a lower bound on the equivalent computational power of the brain.
69
ANN’s: an Operational View
Input S
ignals yk
Output
Signal
wk3
wk1
wk2
wk4
Neuron xk
x1
x2
x3
x4
bk
Summing Junction
Activation Function
)(
)(0
****
:
1
1
1
44332211
THRESHOLDnetifnety
THRESHOLDnetify
functionidentitynegativenon
netyfunctionidentity
outputNeuronal
xwxwxwxwnet
sumweightedorinputNet
70
ANN’s: an Operational
View
A schematic for an 'electronic' neuron
ykwk3
wk1
wk2
wk4
Neuron xkx1
x2
x3
x4bk
Input Signals
Output
Signal
Summing Junction
Activation Function
71
Biological and Artificial NN’s
Entity Biological Neural Networks
Artificial Neural Networks
Processing Units Neurons Network Nodes
Input Dendrites(Dendrites may form synapses
onto other dendrites)
Network Arcs(No interconnection
between arcs)
Output Axons or Processes(Axons may form synapses onto
other axons)
Network Arcs(No interconnection
between arcs)
Inter-linkage Synaptic Contact (Chemical and Electrical)
Node to Node via Arcs
Plastic Connections Weighted Connections Matrix
72
Biological and Artificial NN’s
Entity Biological Neural Networks
Artificial Neural Networks
Output Dendrites bring inputs from different locations: so does the brain wait for
all the inputs and then start up the summing
exercise or does it perform many different intermediate
computations?
All inputs arrive instantaneously and are summed up in the same
computational cycle: distance (or location)
between neuronal nodes is not an issue.
73
The McCulloch-Pitts Network
. McCulloch and Pitts demonstrated that any logical function can be duplicated by some network of all-or-none neurons referred to as an artificial neural network (ANN).
Thus, an artificial neuron can be embedded into a network in such a manner as to fire selectively in response to any given spatial temporal array of firings of other neurons in the ANN.
Artificial Neural Networks for Real Neuroscientists: Khurshid Ahmad, Trinity College, 28 Nov 2006
74
The McCulloch-Pitts
Network
Consider a McCulloch-Pitts network which can act as a minimal model of the sensation of heat from holding a cold object to the skin and then removing it or leaving it on permanently.
Each cell has a threshold of TWO, hence fires whenever it receives two excitatory (+) and no inhibitory (-) signals from other cells at a previous time.
Artificial Neural Networks for Real Neuroscientists: Khurshid Ahmad, Trinity College, 28 Nov 2006
75
The McCulloch-Pitts
Network
1
3
A
B
42
-
+
+
+ +
++
+ +
+
+
Heat
Receptors Cold
Hot
Cold
Heat Sensing Network
76
The McCulloch-Pitts
Network
1
3
A
B
42
-
+
+
+ +
++
+ +
+
+
Heat
Receptors Cold
Hot
Cold
Heat Sensing Network
Time Cell 1 Cell 2 Cell a Cell b Cell 3 Cell 4
INPUT INPUT HIDDEN HIDDEN OUTPUT OUTPUT
1 No Yes No No No No
2 No No Yes No No No
3 No No No Yes No No
4 No No No No Yes No
Truth tables of the firing neurons when the cold object contacts the skin and is then removed
77
The McCulloch-Pitts
Network
Heat Sensing Network
‘Feel hot’/’Feel cold’ neurons show how to create OUTPUT UNIT RESPONSE to given INPUTS that depend ONLY on the previous values. This is known as a TEMPORAL CONTRAST ENHANCEMENT.
The absence or presence of a stimulus in the PREVIOUS time cycle plays a major role here.
The McCulloch-Pitts Network demonstrates how this ENHANCEMENT can be simulated using an ALL-OR-NONE Network.
78
Indexing and AnnotationIn multi-modal information processing, convergence of modalities is critical.
Surveillance experts become experts because they are very good at the convergence.
Surveillance experts LEARN how to mix modalities
From Alex Meredith, Virginia Commonwealth University, Virginia, USA
MultisensoryEnhancement
Cross-modalSuppression
Cross-modalFacilitation
Cross-modalFacilitationInhibition-Dependent
79
词图 CITU (C2) SystemA multi-modal image annotation and retrieval system that can learn to annotate images using artificial neural network systems
80
词图 CITU (C2) System
Manual annotation
Cross modal learning
Database
Image analysis
Image pre-processing
Image segmentati
on
Feature extraction
Language processing
Frequency analysis
Collocations
Feature extraction
Image content
Image feature
Free text
Linguistic feature
Image and
linguistic Features
Cross modal
associations
Image content and free
text
81
Ontogenesis of numerosity
An unsupervised multinet alternative: Simulating Fechners’
Law
Ahmad K., Casey, M. & Bale, T. (2002). Connectionist Simulation of Quantification Skills. Connection Science, vol. 14(3), pp. 165-201.
GatingNetwork
CardinalNumber
Response
“Three”
VisualScene
SubitisingExpert
CountingExpert
82
Ontogenesis of numerosity
An unsupervised multinet alternative: Simulating Fechners’
Law
Ahmad K., Casey, M. & Bale, T. (2002). Connectionist Simulation of Quantification Skills. Connection Science, vol. 14(3), pp. 165-201.
CardinalNumber
VisualScene
648-dVector
15-d to36-d
Vector
Translation-InvariantWeight-Sharing
Network
72-15
Scale-InvariantSecond-Order
Network
648:72
72-dVector
MappingModule
MagnitudeKohonen SOM
36-dVector
Bi-directionalHebbianNetwork
64-dVector
36-36 x 1 36-64 16-8 x 8
VerbalKohonen SOM
MagnitudeRepresentation
64-dVector
83
Simulation of Aphasic Naming
John Wright and Khurshid Ahmad (1997). ‘Connectionist Simulation of Aphasic Naming’. BRAIN AND LANGUAGE Vol. 59, pp 367–389 (1997)
A modular connectionist architecture was developed, in which semantic-lexical and phonological knowledge are instantiated using self-organising Kohonen maps, while connections between them are implemented using Hebbian networks; a linear connectionist network (Madaline) is used to simulate non-word repetition. The Hebbian connections were lesioned in order to reproduce the patient’s naming errors.
84
Ontogenesis of numerosity
One
Tw o ThreeFour
Five
Six
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Kohonen Layer Node Number
Avera
ge A
cti
vati
on
An unsupervised multinet alternative: Simulating Fechners’
Law
Ahmad K., Casey, M. & Bale, T. (2002). Connectionist Simulation of Quantification Skills. Connection Science, vol. 14(3), pp. 165-201.
85
Ontogenesis of multi-modal behaviour
Jacob Martin, Alex M. Meredith, and Khurshid Ahmad. (2009) ‘Modeling multisensory enhancement with self-organizing maps’. Frontiers in Computational Neuroscience (June 2009), Volume 3 (Article 8), pp 1-10.
86
Sentiment analysis of Irish Newspaper Texts and the Irish Stock Exchange Index.
0.00
0.02
0.04
0.06
0.08
0.10
0.12
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Years
Vol
atili
ty
ISEQ
Negative
Positive
87
Computers and Brain: A neuroscience perspective
Alan Turing (1950) ‘Computer Machinery and Intelligence’. Mind Vol. LIX (No. 2236), pp 433-460.
“Professor Jefferson's Lister Oration for 1949, from which I quote.
"Not until a machine can write a sonnet or compose a concerto because of thoughts and emotions felt, and not by the chance fall of symbols, could we agree that machine equals brain-that is, not only write it but know that it had written it.
No mechanism could feel (and not merely artificially signal, an easy contrivance) pleasure at its successes, grief when its valves fuse, be warmed by flattery, be made miserable by its mistakes, be charmed by sex, be angry or depressed when it cannot get what it wants."