more from this thinker     |     more from this text

Single Idea 19737

[filed under theme 14. Science / A. Basis of Science / 4. Prediction ]

Full Idea

If we can train a system for prediction, it can essentially infer the structure of the world it's looking at by doing this prediction.

Gist of Idea

A system can infer the structure of the world by making predictions about it

Source

New Scientist writers (New Scientist articles [2013], 2015.12.12)

Book Ref

-: 'New Scientist magazine' [ 2013], p.29

A Reaction

[AI expert] This seems to be powerful support for the centrality of mathematical laws of nature in achieving understanding of the world. We may downplay the 'mere' ability to predict, but this idea says that the rewards of prediction are very great.

The 18 ideas with the same theme [predicting events as support for a theory]:

12738 Successful prediction shows proficiency in nature [Leibniz]
22179 Explanatory facts also predict, and predictive facts also explain [Hempel, by Okasha]
13054 Correlations can provide predictions, but only causes can give explanations [Salmon]
18960 Most predictions are uninteresting, and are only sought in order to confirm a theory [Putnam]
8057 Unpredictability doesn't entail inexplicability, and predictability doesn't entail explicability [MacIntyre]
17087 The 'symmetry thesis' says explanation and prediction only differ pragmatically [Ruben]
16856 It is more impressive that relativity predicted Mercury's orbit than if it had accommodated it [Lipton]
16857 Predictions are best for finding explanations, because mere accommodations can be fudged [Lipton]
5340 Explanation does not entail prediction [Flanagan]
17374 The possibility of prediction rests on determinism [Dupré]
4807 A good barometer will predict a storm, but not explain it [Psillos]
4808 If we say where Mars was two months ago, we offer an explanation without a prediction [Psillos]
6757 Explanation predicts after the event; prediction explains before the event [Bird]
18698 Predictions give the 'content' of theories, which can then be 'equivalent' or 'adequate' [Button]
14915 The theory of evolution was accepted because it explained, not because of its predictions [Ladyman/Ross]
14916 What matters is whether a theory can predict - not whether it actually does so [Ladyman/Ross]
17405 If a theory can be fudged, so can observations [Scerri]
19737 A system can infer the structure of the world by making predictions about it [New Sci.]