This document summarizes a neural model of human image categorization. It describes using a leaky integrate-and-fire neuron model and deep autoencoder with circular convolution to represent classes of visual objects. It also summarizes two classic studies on visual categorization - Posner & Keele's prototype theory study from 1968 and Regehr & Brooks' exemplar theory study from 1993. The model is able to account for human performance in both studies by representing object classes with prototypes or exemplars.
DSPy a system for AI to Write Prompts and Do Fine Tuning
Neural Model of Human Image Categorization - LIF Neuron, Deep Autoencoder & Prototype/Exemplar Theories
1. Summary of
A Neural Model of
Human Image Categorization
Methodology of Cognitive Science
Jin Hwa Kim
Cognitive Science Program
Seoul National University
2. What We Will See
1. Computational Neural Model
-
Leaky integrate-and-fire (LIF) neuron model
-
Deep autoencoder
-
Circular convolution
2. How classes of visual objects are represented in the brain?
-
Prototype-based (Posner & Keele, 1968)
-
Exemplar-based (Regehr & Brooks, 1993)
3. LIF Neuron Model
Leaky integrate-and-fire (LIF) neuron model
-
One of biological neron models (spiking neuron model)
[Gerstner and Kistler, 2002]
6. Deep Autoencoder
Specialized neural network
-
Try to make the output be the same as the input in a
network with a central bottleneck
output vector
decoding
weights
semantic pointer
code
encoding
weights
input vector
7. Deep Autoencoder
Solving optimization problem
-
Use unsupervised layer-by-layer pre-training.
-
LIF instead of RBM
W1
W2
W3
784 ! 1000 ! 500 ! 250 W4
W1T
T
W2
T
W3
784 " 1000 " 500 " 250
30 linear units
T
W4
We train a stack of 4 RBM s and then unroll them.
Then we fine-tune with gentle backprop.
[Hinton & Salakhutdinov, 2006]
8. Circular Convolution
Store semantic pointers
-
Holographic reduced representations using compositional
distributed representation
circular
convolution operator
Categorization process
[Plate, 2003]
10. Posner & Keele, 1968
Prototype theory
-
It was designed to test whether human subjects are
learning about class prototypes when they only ever see
distorted examples.
Figure 3: Sample stimuli for Experiment 1, modelling a classic study by Posner & Keele (1968).
The dot patterns are created by distorting three randomly drawn prototype images (left) with
low (centre) and high (right) levels of noise. Subjects are trained to classify a set of twelve highdistortion patterns and tested without feedback on the same prototypes at different distortion
levels.
11. Posner & Keele, 1968
Results
Figure 4: Comparison of human and model performance for Experiment 1. The model is able to
account for human results when presented with the schema, low distortion (5), and high
distortion (7) patterns. Occasional random errors by human subjects may explain the
discrepancy on training examples. Error bars indicate 95% confidence intervals. Human data
from Posner & Keele (1968).
12. Regehr & Brooks, 1993
Exemplar theory
-
Analytic vs. Perceptual similarity
Figure 5: Sample stimuli for Experiment 2, modelling experiment 1C of Regehr and Brooks (1993). (Left) Images are
composed of interchangeable (composite) feature manifestations. (Right) Images expressing the same attributes are
drawn in a more coherent (individuated) style. Regehr & Brooks (1993) drew a distinction between good transfer and bad
transfer test stimuli. A test stimulus is a good transfer case when the addition or removal of spots matches a training case
with the same label, and a bad transfer case if adding or removing spots matches a training case with the opposite label.
(Adapted from Regehr & Brooks (1993) Figure 2A).
13. Regehr & Brooks, 1993
Results
Figure 6: Comparison of human and model performance for Experiment 2. Our model
accounts for the key difference in human performance on the good transfer (GT) versus bad
transfer (BT) pairs for the individuated stimuli. Error bars indicate 95% confidence intervals.
Human data from Regehr & Brooks (1993).