Atlas record: https://www.aitimeline.dev/en/events/evt-0215/
Milestone · March 2009
The unreasonable effectiveness of data
Three Google researchers argued that for language a simpler model on enormous data beats an elegant model on little, so effort belongs in the data.
Why it matters
A position the industry had reached empirically was finally stated as a principle, and the field began measuring itself by corpus size.
The title echoes Wigner's essay on the unreasonable effectiveness of mathematics. The authors argue that language is inherently complex and perhaps doomed to theories without the elegance of physics equations, so one should embrace the complexity and rely on data rather than seek an elegant theory. The thesis proved both right and one-sided: scale gave a great deal, but the move from n-grams to transformers was a change of model.
Event record
- Event date
- March 2009
- Timeline date
- Event date
- Verification
- Sources gathered automatically · September 17, 2026
- Lines
- ID
- evt-0215
The March-April 2009 issue of IEEE Intelligent Systems.
Related events
Earlier
- Enables A hole in a picture filled from other photographs 2007links to this record
"The Unreasonable Effectiveness of Data" cites this very work: with thousands of photos the results were poor, with millions the same algorithm worked well.
- Related MapReduce 2004links to this record
The practical way of processing the volumes the data-effectiveness thesis rests on.
- Extends Data outweighs the algorithm 2001
It develops the Banko and Brill result into a general position on data and models.
Later