Utility Functions (Agent Preferences)

in #economics5 years ago

Utility Functions and Agent Preferences

In the past I discussed the concepts of happiness maximization and computational kindness. Human beings like to have relationships (obviously right? In my opinion the concept of the utility function is one of the more practical concepts in the theory of economics because for lack of a better word it "mathematizes" social interaction. Social exchange theory in fact applies the concepts from economics, behaviorism, and shows that these concepts are legitimate in a sociological or anthropological context.

The concept of the utility function also make a lot of sense when discussing a platform like Steem and I'm actually surprised I do not see it discussed all that much. Content has value on the basis of utility theory in economics. That said at least one poster on Steem has posted on these topics @spectrumecons. In my opinion his posts are underappreciated and I think more attention should go toward some of his work.

Some examples of his work include:

To understand what a utility function is, I will simply reveal what it does. A utility function is a mathematical model of the preferences of an entity. The math formula representing a utility function is: U(x1,x2,x3…).

To bridge a gap in understanding we can now connect the concept of a utility function (economics and mathematics) to the programming world by imagining how it relates to preference networks. Conditional preference networks in specific are a language for the structured representation of ordinal utility. For those familiar with VC theory, you can see also know machine learning is implementable using CP-Nets such that the learning mechanism can attempt to identify the preferences of the target entity. Finally for programmers, data structures exist which allow for the convenient implementation of preference databases (such as trees or graphs).

On Steem Proof of Brain is an algorithm of collaborative filtering which organizes content in accordance with the collective utility function of the social network stakeholders. Meaning the collective preferences of the stakeholders is what determines the worth of a piece of content. I might not have explained it perfectly, and anyone with a better grasp on economic theory is free to correct my interpretation.

References

  1. Koriche, F., & Zanuttinib, B. (2010). Learning conditional preference networks. Artificial Intelligence, 174, 685-703.
Sort:  

To listen to the audio version of this article click on the play image.

Brought to you by @tts. If you find it useful please consider upvoting this reply.

Coin Marketplace

STEEM 0.35
TRX 0.12
JST 0.040
BTC 70601.11
ETH 3576.21
USDT 1.00
SBD 4.78