AI ki Kahani, Part 3

Do ladke ek hi school mein padhe the. Ek ne seekhne wali machine banayi. Doosre ne ek kitaab likh kar dikhaya ki wo machine kahaan haar jaati hai.

Part 3 mein: 1969 mein MIT ke Marvin Minsky aur Seymour Papert ne kitaab likhi, naam tha "Perceptrons". Unhone ganit se saabit kiya ki ek layer wala perceptron kuch bahut aasaan sawaal bhi hal nahi kar sakta. Sabse aasaan sawaal: XOR. Seedhi wali light socho. Light tab jale jab sirf ek switch ON ho. Ek switch se jod threshold paar karta hai, toh dono ON hone par jod aur bada ho jaata hai aur light galat jal jaati hai. Perceptron haan aur na ke beech sirf ek seedhi lakeer kheench sakta hai. Iske baad neural networks par paisa aur bharosa dono ghat gaye. Is daur ko AI winter kehte hain. Hal tab bhi pata tha: beech mein ek aur layer. Mushkil ye thi ki us layer ko sikhaayein kaise.

Aaj ye karo: apni ek atki hui problem likho aur use do chhote sawaalon mein todo. Phir har sawaal ka jawab alag se dhoondho.