Difference between revisions of "Inductive bias"

From Lesswrongwiki
Jump to: navigation, search
(See Also)
m (byline removal)
 
(6 intermediate revisions by 3 users not shown)
Line 1: Line 1:
 
{{wikilink}}
 
{{wikilink}}
 
 
{{Quote|
 
{{Quote|
 
The inductive bias of a learning algorithm is the set of assumptions that the learner uses to predict outputs given inputs that it has not encountered
 
The inductive bias of a learning algorithm is the set of assumptions that the learner uses to predict outputs given inputs that it has not encountered
|(Mitchell, 1980)|[http://citeseer.ist.psu.edu/mitchell80need.html The need for biases in learning generalizations]}}
+
|Tom Mitchell<ref>
 +
{{cite journal
 +
|author = Tom M. Mitchell
 +
|year = 1980
 +
|title = The need for biases in learning generalizations
 +
|url = http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.120.4179
 +
}}</ref>}}
 +
 
 +
'''Inductive bias''' refers to your suspicion that if the sun has risen for the last billion days in a row, then it may rise tomorrow as well.  Since it is [[logically possible]] that the laws of physics will arbitrarily cease to work and that the sun will ''not'' rise tomorrow, coming to this conclusion requires an inductively biased [[prior]].
 +
 
 +
This sort of bias is not a bad thing - without "inductive bias" you can't draw any conclusion at all from the data.  It's just a different technical meaning attached to the same word.
  
In [[Bayesian]] framework, inductive bias is encoded in the [[prior distribution]].
+
==Primary post==
  
==See Also==
+
*[http://lesswrong.com/lw/hg/inductive_bias/ "Inductive Bias"]
 +
 
 +
==Footnotes==
 +
<references/>
 +
 
 +
==See also==
 +
 
 +
*[[Superexponential conceptspace]]
 
*[[Prior distribution]]
 
*[[Prior distribution]]
*[[Bias (statistics)]]
+
*[[Statistical bias]]
 
+
*[[Cognitive bias]]
==References==
 
<!-- Always keep this header if there is at least one reference
 
    Delete or add sections below as necessary -->
 
  
=====Overcoming Bias Articles=====
+
[[Category:Machine learning]]
<!-- For related Overcoming Bias articles,
 
    formatted as a list with each entry as "/Title/ by /Author/ -->
 
*[http://www.overcomingbias.com/2007/04/inductive_bias.html "Inductive Bias"] by [[Eliezer Yudkowsky]]
 
{{stub}}
 

Latest revision as of 06:48, 18 November 2009

Smallwikipedialogo.png
Wikipedia has an article about

The inductive bias of a learning algorithm is the set of assumptions that the learner uses to predict outputs given inputs that it has not encountered

Tom Mitchell[1]

Inductive bias refers to your suspicion that if the sun has risen for the last billion days in a row, then it may rise tomorrow as well. Since it is logically possible that the laws of physics will arbitrarily cease to work and that the sun will not rise tomorrow, coming to this conclusion requires an inductively biased prior.

This sort of bias is not a bad thing - without "inductive bias" you can't draw any conclusion at all from the data. It's just a different technical meaning attached to the same word.

Primary post

Footnotes

See also