The learning idea was already there

Learning did not suddenly appear after rule-based AI ran into trouble. The idea had been there near the beginning.

In 1957, only a year after the Dartmouth workshop, a psychologist named described a system called the .

Researchers such as Newell and Simon were building programs from symbols, written rules, and search. Rosenblatt asked a different question: how could a machine learn to recognise something from examples?

He looked to the brain for inspiration. Instead of giving the machine a written rule for every pattern, he proposed connections whose influence could change after each example.

The two approaches sometimes overlapped, but they placed their bets in different places. One tried to write down the reasoning. The other let examples reshape the machine.

To understand Rosenblatt's machine, we first need the simple artificial neuron it was built from.

The learning idea was already there

Learning did not suddenly appear after rule-based AI ran into trouble. The idea had been there near the beginning.

In 1957, only a year after the Dartmouth workshop, a psychologist named described a system called the .

He looked to the brain for inspiration. Instead of giving the machine a written rule for every pattern, he proposed connections whose influence could change after each example.

The two approaches sometimes overlapped, but they placed their bets in different places. One tried to write down the reasoning. The other let examples reshape the machine.

To understand Rosenblatt's machine, we first need the simple artificial neuron it was built from.