Science & tech
The “Artificial intelligence” quiz — Science & tech. It has 25 questions in five categories: AI history, Training modes, Deep models, Language & attention, Search & decisions. Questions get harder as the points go up, from 100 to 500.
The 100- and 200-point questions are open here. The rest show only their first words, and there are no answers: all of that is in the game itself.
Who is pictured here as a teenager, the man who proposed the “imitation game” in a 1950 paper on machine intelligence?
Who coined the term “artificial intelligence” in the proposal for the Dartmouth summer project?
Who created the perceptron…
In the gameWhat 1970s Stanford expert…
In the gameWho introduced the Chinese…
In the gameWhat training paradigm uses examples paired with known target labels?
Which paradigm finds structure such as clusters in data that has no target labels?
Which approach trains an…
In the gameWhat problem occurs when…
In the gameWhat method adds a…
In the gameWhich activation function returns max(0, x)?
Which algorithm sends loss gradients backward through layers so weights can be updated?
Which architecture slides learned…
In the gameWhich recurrent architecture uses…
In the gameWhich layer standardizes activations…
In the gameWhich 2017 architecture replaced recurrence and convolutions with attention mechanisms?
Which modeling approach predicts the next token from previous tokens?
What mechanism prevents a…
In the gameWhich attention…
In the gameWhich mechanism projects Q…
In the gameWhich search algorithm evaluates a node with f(n)=g(n)+h(n)?
Which game-tree rule assumes both players act optimally in a two-player zero-sum game?
Which pruning method skips…
In the gameWhich property says the…
In the gameWhich named recursion defines…
In the gameAll 25 questions and the answers are in the game. Invite your class or friends with a link — guests need no account.