AI History Battle

Whose methods actually
solve the problem?

Every coin is a real figure from the history of statistics, computing, and machine learning. Tap someone else's coin and you both get a real problem from the field's past — the winner is whoever's actual toolkit fits it. Fisher beats Hinton on eight field plots. Hinton buries Fisher on sixty thousand digits.

Tap your own coin against your phone to begin.
Hirotugu Akaike Richard Sutton Donald Hebb Herbert Simon Lek-Heng Lim Kiyosi Ito Martin Wainwright Marvin Minsky David Blackwell Richard Karp John Hopfield Jon Kleinberg Richard Bellman Trevor Hastie
64playing
163battles
152figures
305problems

Leading

all awards →
1 Michael I. Jordan
stat-learning
11–1
2 David Blei
stat-learning
9–1
3 John Santerre
deep-modern
9–19

Most wins is only one way to win — there are nine boards, and a weak figure can still take Overachiever.

Just now

Jon Kleinberg beat John Santerre
Drive through the intersection
Terence Tao beat John Santerre
Program chess before the computer exists
Larry Wasserman beat Frank Rosenblatt
One test or twenty?
John Santerre beat Michael Mahoney
The memory that completes the pattern
Martin Wainwright beat Michael I. Jordan
Just look at the neighbors
Linus Torvalds beat Karl Pearson
The robot in the warehouse

Most fought over

The memory that completes the pattern 3 battles

perception — see who wins it and why

Explore

152The rosterBayes to the transformer 305The problemsEvery question in play 9AwardsMore than one way to win The graphWho has met whom

Every result comes from 15,324 precomputed judgments about whose methods apply where. Nothing is decided at random, and the explanations are the point — read why you won.