Congratulations to Alex Anderson for graduating with his Ph.D.! His thesis made an important contribution to our understanding of the role of fixational eye movements in perception, demonstrating the feasibility of neural computations that can simultaneously estimate form and motion from retinal drift, thus providing a first account for how and why visual acuity improves with a moving retina. He also developed a recurrent neural network model for predicting the location of the eye from initial onset of an eye movement, and he provided a theoretical account for why neural network activations and weights may be binarized with only nominal effects on performance, thus resulting in a dramatic reduction in compute time and power consumption.
Alex will be starting his new job as a research scientist at WaveOne on July 2nd, where he will be creating context-adaptive image compression algorithms. Our loss, their gain, and we pity their competitors. We’ll miss you Alex!