Neural Networks And Artificial Brains

Question 1

What does our brain consist of?

Answer 1

Highly interconnected neuronal networks

Question 2

How many neurons are there?

Answer 2

100 billion

Question 3

What is the volume of the brain?

Answer 3

1.5 litres

Question 4

How much power does the brain consume?

Answer 4

10W

Question 5

What is 50-80% of total energy consumption in the brain used for?

Answer 5

Conduction of action potentials along nerve fibres and in synaptic transmission

Question 6

What is the other 20-50% of the brain energy consumption used for?

Answer 6

Manufacturing and maintenance

Question 7

How do brain transmit signals so quickly despite being so slow?

Answer 7

It is constructed as highly parallel networks, most neurons connect directly to many thousands of others

Question 8

How does brain form so many connections?

Answer 8

Brain exploits it’s three-dimensional volume to pack everything by bending sheets of cells into folds and weaving connections closely together into bundles

Question 9

How do silicon neurons overcome the limitations of the 2D nature of chips and circuit boards?

Answer 9

Exporting the very high speed of conventional electronics where the same wire can carry many different messages along the same wire

Question 10

what has neurally-inspired engineers adopted to reduce power but increase speed?

Answer 10

Using analogue rather than digital coding

Question 11

How do analogue circuits work?

Answer 11

Instead of coding in 0s and 1s, it codes in continuous changes in voltages

Question 12

How do neurons compute and make decisions?

Answer 12

They transmit impulses down axons to communicate the answer to target neurons

Question 13

What is a disadvantage about spike coding?

Answer 13

It is energetically costly

Question 14

How do efficient coding maximise the information represented in a pattern of spikes?

Answer 14

By reducing redundancy

Question 15

What is an advantage of sparse coding?

Answer 15

Increasing energy efficiency

Question 16

What is sparse coding?

Answer 16

Using as small number of active neurons as possible

Question 17

What is a simple artificial version of a biological network built?

Answer 17

A silicon retina that captures light and adapts its output automatically to changes in overall lighting condition

Question 18

How does a silicon retina that capture light work?

Answer 18

It connects to two silicon neurons like real neurons in the visual cortex and have the job of extracting information about the angles of lines and contrast boundaries in retinal image

Question 19

What are neurons in silicon retina prototype called?

Answer 19

Integrate and fire neurons

Question 20

How do integrate and fire neurons get their name?

Answer 20

They add up the weighted inputs, coded as voltages that are arriving at their synapses and only fire an action potential if the voltage reaches a set threshold

Question 21

What are silicon neurons like?

Answer 21

They are built of transistors and operate in their sub threshold range and act like cell membranes of real neurons, additional transistors provide active conductance to emulate the voltage and time-dependent current flows of real ion channels

Question 22

What does artificial neural networks consistent of?

Answer 22

They consist of a number of simple processing units that are highly connected in a network

Question 23

What is the simplest form of artificial neural network?

Answer 23

Feedforward associator

Question 24

What does a feedforward associator have?

Answer 24

Layers of interconnected input and output units

Question 25

How is associative memory encoded by?

Answer 25

Modifying the strengths of connections between layers such that when an input pattern is presented, the stored pattern associated with that pattern is retrieved

Question 26

What is a more complex artificial neural network?

Answer 26

Recurrent neural net

Question 27

What does a recurrent neural net consist of?

Answer 27

A single layer where every unit is interconnected and all units act as input and ouput

Question 28

What does the design of recurrent neural net allows it to do?

Answer 28

Store patterns rather than merely pairs of items

Question 29

How to decode autoassociative networks?

Answer 29

Recursive search for a stored pattern

Question 30

In a network of 1000 units, how many patterns can be retrieved before errors in the retrieval patterns become too large?

Answer 30

150

Question 31

What is the similarity between artificial neural networks and brain?

Answer 31

The way they store and process information

Question 32

What kind of storage does artificial neural network have?

Answer 32

Content-addressable storage

Question 33

How are information stored in artificial neural networks?

Answer 33

Information is stored in the weights of connections, the same way that synapses change their strength during learning

Question 34

What are the learning rules that change artificial neural networks?

Answer 34

They train it by modifying the strength of connections between neurons by taking the output of the network to a given input pattern and comparing it with the desired pattern, the difference which is the error is then used to adjust the weights of the connections to achieve a closer output to the desired one and gradually reduces the error to the minimum

Question 35

What is the problem of artificial neural networks?

Answer 35

It is simulated mathematically on digital computers and this takes time so artificial neural networks cannot operate in real time