Lecture 10 : Perceptron Logic OR gate

[drudru]

Hello,

In Lecture 10 - slides 9-11:

Shouldn't the bias be -0.5 (some number < 0) for the OR() perceptron?
aka w = (-0.5,1,1). Otherwise an input of (0,0) on (1.5,1,1) would be at 1.5 or True.

AND() is w = (-1.5,1,1), and that is consistent with SIGN as the threshold.

I searched the slides and I couldn't find a definition of the perceptron threshold function other than SIGN.

Dru

[yaser]

Originally Posted by drudru

Hello,

In Lecture 10 - slides 9-11:

Shouldn't the bias be -0.5 (some number < 0) for the OR() perceptron?
aka w = (-0.5,1,1). Otherwise an input of (0,0) on (1.5,1,1) would be at 1.5 or True.

AND() is w = (-1.5,1,1), and that is consistent with SIGN as the threshold.

I searched the slides and I couldn't find a definition of the perceptron threshold function other than SIGN.

The binary convention we use is rather than . With this convention, the above values work out.

[drudru]

Ah, thank you.