Frame Problem

FRAME PROBLEMAnkit Awasthi

R Lakshminarayan Bharadwaj

Sudhanshu Chowdhary

Vinay Chataraju

Vipul Arora

Philosophy 141

Guide:Prof. Prashant Bagad

FRAME PROBLEMORIGIN AND RELEVANCE

Origin : Artificial Intelligence

McCarthy and Hayes (1969).

Matrix, Terminator, Minority report etc. (????)

Dennett(1978), Fodor(1987) …

DEFINITION

If we had a number of axioms to be performed in sequence we would have quite a number of conditions to write down that certain actions do not change the values of certain properties(fluents). In fact with ‘n’ actions and ‘m’ fluents , we might have to write down ‘mn’ such conditions.

The chief motivation for frame problem is that most things don't change when an action is performed or an event occurs.

BASIC TERMS

Fluents: Properties whose values can change as a result of an action.

Effect Axioms: Axioms that state the effects of an action.

Frame Axioms: Axioms that state the non-effects of an action.

State: The collection of fluent values under consideration.

AN EXAMPLE

Fluents:House_color() Initial conditions ( t=0):Furniture_position() House_color(Red)

Furniture_position(#)Actions:Paint(house,color): changes color of the house to ‘color’Rearrange(furniture,position): change the position of the furniture to ‘position’

EXAMPLE ( C O N T D . )

At t=0, take action

Paint(House,Yellow)

At t=1,

What is the position of the furniture???

EXAMPLE ( C O N T D. )

The problem becomes even worse if we allow concurrent actions.

In this case frame axioms are in fact wrong .

An exception needed??

A infinite regression………..

EPISTEMOLOGICAL ASPECTS

How “a cognitive creature … with many beliefs about the world” can update those beliefs when it performs an act so that they remain “roughly faithful to the world”?

Given the action of picking up a tea cup how does a robot update your beliefs about the world (ambient temperature, current date, etc.)

COMPUTATIONAL ASPECT OF FRAME PROBLEM

This is the question of how to compute the consequences of an action without the computation having to range over the action's non-effects.

‘Sleeping dog’ strategy .

Appeal to relevance

BOT BATTERY EXAMPLE

Problem statement: program a bot to search and obtain batteries if present.

What if batteries are tied with a thread ????

What if you take every possibility into account, but in the process of evaluation a bomb explodes??

COMMON SENSE FRAME PROBLEM

It is based on the notion that it might not possible to axiomatise even a significant portion of the real world .

Common sense reasoning deals with ill defined concepts.

The problem of ‘kinds’: precisely what constitutes a particular class.

The entire commonsense world seems populated with categories of things which we find immensely useful to think about in planning our daily getting around and yet which defy ‘precise definition’ .

APPROACHES TO SOLVING FRAME PROBLEM

A naïve approach could be to assign probabilities to events.

Resorting to Metaphysical concepts like “common sense law of intertia”.(Does it work??)

MOOT QUESTIONS!!

QUESTIONS … !