Usage of variance instead of stddev in `select` of LinUCB

[mayank010698]

Shouldn't this be np.sqrt(sig) ?

EE-Net-ICLR-2022/baselines/LinUCB.py

Line 18 in 08d6ed7

r = np.dot(context, self.mu) + np.sqrt(self.lamdba * self.nu) * sig