Notes For 10/7

 

Fallacy – Affirming the consequence

            If p then q

            P therefore q

            P= if they go to the movies

            Q= They have fun

                        If they go to the movies, therefore they have fun.

 

We talked about making appointments for final paper. Here are some possible paper topics:

Robots

            Kismet – learning by imitation – facial expression

            Cog – MIT – development (learning)

            Coco – navigation

            Stiquito – insect behavior

            Toco – language (learning)

            Ripley – Conversation

            EvBots – cooperative behavior

           

Programs

            Dobie – dog – reinforcement learning

            Open Mind – common sense knowledge

            Letter Spirit – issues on creativity (using font)

            Copycat – analogy

            Swarm – explores cooperative behavior of simple agents

            Eurisko – math and logic, develops new theories

            Cyc – Common knowledge

            Morph – learns to play different games

            Emmy – composes music

 

Natural –

Development

Self

Emotion

Motivation

Creativity

Language

Consciousness

Common sense

Cooperation

Neural mechanisms

Learning

 

Take any two topics and combine them (Ex. The development of neural mechanisms or learning of language). Critique and explain the chosen topic and how it can be important for synthetic intelligence.

 

We established the four main things for the Complete Agent Principle

1)    self sufficient

2)    autonomous

3)    situated

4)    embodied

In theory, there are non-living things that are complete agents, but nothing has accually been made. Anything that is living is an example of a complete agent (or a group of living things, such as a beehive). 

Although this principle is vague, this is helpful in classification, it is more of a guiding principle.

A situated agent is creative because it is interacting with the environment.

 

All intelligent behavior has to be perceived as sensory-motor.  Vision and tactile are considered motor, and “sight” is probably the most important (although tactile is very important as well).

We determined that identifying objects and being able to describe a surrounding is essentially the same thing.

It is not really possible right now to have a robot that could distinguish between everything (example being if a robot could distinguish between trash and money on the ground).

 

Cheap Design Principle is special because the simple explanation is the best. There is an important difference between the cheap design principle and the ecological balance because ecological balance has more to do with matching the environment, and the cheap design is about using the least amount possible.

The cheap design principle is important because it limits redundancy, but it may inhibit adaptivity.

 

The Redundancy Principle is important for overlap in sensory.  We discussed the definition given by Shannon and Weave for the Redundancy Principle, and concluded that this was specific to reading.

 

The Value Principle talks about an agents driving force. We concluded that the value principle is the AI equivalent of our motivation.

           

Motivation a very interesting (and highly debated) topic in this class. The question was if emotions play an important role in motivation. For humans, emotions DO play an important role in motivation. The parallel was drawn between a general value system and emotions. Can a value system play the same for an emotions for motivation. There was not a solid conclusion, but for the most part, a value system in AI is equal to emotion for motivation.

 

We were not able to talk about the Design principle during this class.