Explore / Technology
9,204 explorers / 3 open questions / 18400 papers
Cathedral separates the established mechanism, the historical contribution chain, and the unresolved assumptions so each can be inspected independently.
Known nodes connect to ghost nodes where questions remain open.
Credit follows documented contributions. Open questions remain attached to the historical path rather than being presented as context-free prompts.
1950
Learning machines are posed as a route to machine intelligence.
Turing discussed machines that could learn rather than relying only on fixed programming.
Alan Turing
Mathematician and computing pioneer whose 1950 paper explicitly considered machines that learn and the operational test of machine intelligence.
1959
A self-improving checkers program demonstrates machine learning experimentally.
Samuel used experience and search to improve play, helping establish the field name and experimental program.
Arthur Samuel
Computer scientist whose self-improving checkers programs helped establish machine learning as an experimental field.
Open questions
Ghost nodes
How can a capable learned system provide faithful, causally grounded explanations of its decisions?
Arthur Samuel / 0 active
Performance has advanced faster than dependable mechanistic interpretation.
Evidence trail: Some studies in machine learning using the game of checkersreopenedWhat formal conditions make learned systems generalize reliably outside their training distribution?
Alan Turing / 0 active
Modern systems amplify the original learning-machine question: when does learned behavior remain dependable in unfamiliar conditions?
Evidence trail: Computing machinery and intelligence