AI ki Kahani, Part 2

July 1958 mein America ke ek akhbaar ne likha ki Navy ki ek machine ek din chalegi, bolegi, dekhegi aur likhegi. Par us machine ki asli khoobi kuch aur thi. Wo khud seekhti thi.

Part 2 mein: Frank Rosenblatt ka Perceptron. Pehle demo mein ek kamre jitne bade IBM computer ko cards dikhaye gaye, jin par baayein ya daayein ek nishaan tha. 50 trials ke baad computer ne khud pehchaanna seekh liya ki nishaan kis taraf hai. Seekhne ka niyam simple tha. Jawab sahi ho toh kuch nahi badalta. Jawab galat ho toh jo inputs us jawab mein shaamil the, unke weights thode badal diye jaate hain. Mark I Perceptron mein 400 light sensors the, aur weights asli knobs the, jinhe galti hone par chhoti electric motors khud ghumati thi.

Aaj ye karo: apni haal ki ek galti likho, aur uske aage wo ek cheez likho jo aap agli baar thoda sa badloge. Sab kuch nahi, sirf ek weight.