
# Natural and Artificial Intelligence

Amazon Thursday Thoughts

Are we close to creating intelligence?

### What is Intelligence?

• Poorly defined.
• My definition: use of information to achieve goals more efficiently.
• Efficiency is defined through use of less resource.
• Automated decision making: we are continuing to progress well.

### The Silicon Factor

• BBC 1, London, 14th September 1980
• Series of three programmes investigating the so-called microelectronics revolution.
• The promise (and perils) of silicon was broadly similar to that for AI today.

### Public Definition

• Doing things that humans do.
• There is a narcissistic element to our understanding of artificial intelligence
• It’s a shifting definition.
• Intelligence is the stuff I can do computers can’t.

### Cybernetics and the Ratio Club

• Ideas came out of the Second World War.
• Researchers explored the use of radar (automated sensing) e.g. Donald MacKay
• Automatic computation for decryption of military codes (automated decision making) e.g. Jack Good and Alan Turing
• Post-war potential for electronic emulation of what had up until then been the preserve of an animallian nervous system.

### The Centrifugal Governor

On Governors, James Clerk Maxwell 1868

$\text{data} + \text{model} \xrightarrow{\text{compute}} \text{prediction}$

### From Model to Decision

 $\text{data} + \text{model} \xrightarrow{\text{compute}} \text{prediction}$

### Artificial vs Natural Systems

• Consider natural intelligence, or natural systems
• Contrast between an artificial system and an natural system.
• The key difference between the two is that artificial systems are designed whereas natural systems are evolved.

### Natural Systems are Evolved

Survival of the fittest

?

### Natural Systems are Evolved

Survival of the fittest

Herbet Spencer, 1864

### Natural Systems are Evolved

Non-survival of the non-fit

### Mistake we Make

• Equate fitness for objective function.
• Assume static environment and known objective.

.

### Engineering Systems Design

• Major component of all Engineering disciplines.

• Details differ: there is a common theme: achieve your objective with the minimal use of resources to do the job.

• This provides efficiency.

• Engineering designer imagines a solution that requires the minimal set of components to achieve the result.

• A water pump has one route through the pump.

### Don’t Fail

• First criterion of a natural intelligence is don’t fail.

• In contrast, mantra for artificial systems is to be more efficient.

• Artificial systems are given a single objective (in machine learning it is encoded in a mathematical function)

• Aim to achieve that objective efficiently.

### Designing out Failure

• Even if we wanted to incorporate don’t fail in some form, it is difficult to design for.

• To design for “don’t fail”, you have to consider every which way in which things can go wrong, if you miss one you fail. These cases are sometimes called corner cases.

### Corners Everywhere

• In an uncontrolled environment, almost everything is a corner.

• It is difficult to imagine everything that can happen.

• Most of our automated systems operate in controlled environments (e.g. a factory, a set of rails.)

### Deployment in Uncontrolled Environments

• Requires a different approach to systems design.
• One that accounts for uncertainty in the environment
• One that is robust to unforeseen circumstances.

### An Intelligent System

Joint work with M. Milo

### An Intelligent System

Joint work with M. Milo

• Need to deal with uncertainty and increase robustness.
• Today, it is easy to make a fool of an artificial intelligent agent.
• Technology needs to address the challenge of the uncertain environment to achieve robust intelligences.
• Successful deployments of intelligent systems are common.
• But they are redefined to be non-intelligent.
• My favourite example is the Centrifugal governor.

### Story

• A man and his dog

### Computers

• A hundred years ago computers were human beings.
• Digital computers originally called automatic computers
• Do we think of such a computer as intelligent?
Are we close to creating intelligence?

1. Current technology is a long way from emulating all aspects of human intelligence: there are a number of technological breakthroughs that remain before we crack the fundamental nature of human intelligence.
1. More controversially, I believe that there are aspects of human intelligence that we will never be able to emulate, a preserve that remains uniquely ours.
 bits/min billions 2000 6 billioncalculations/s ~100 a billion a billion embodiment 20 minutes 5 billion years 15 trillion years

### Conclusion I

• We are a long way from emulating human intelligence, animal intelligence, animal motion.
• The objectives of cybernetics still have not been reached.
• The robustness of natural systems is outside the scope of our current design methodologies.

### Conclusion II

• There is something quintisential about the human experience.
• We are co-evolved to view the world in a certain way to enable collaboration.
• Our consciousness is a consequence of our limitations. Our locked-in intelligence.