Hey, I'm David Nicholson. I am a neuroscientist at Emory University in Atlanta, Georgia. I work with Astrid Prinz in the Biology department. We are developing machine learning algorithms that learn continuously, based on neuroscience, as part of a DARPA grant.
I also work in applied machine learning, most recently in collaboration with Yarden Cohen (twitter, github). We are applying neural networks to segment and annotate the song of songbirds, which in neuroscience serve as a model system to understand how the brain learns and produces speech and similar motor skills, like playing guitar or swinging a bat to hit a baseball. In addition I maintain a software library, hybrid-vocal-classifier, that automates annotation of birdsong for behavioral experiments. These projects began during my graduate studies in Sam Sober's lab. My dissertation work in the Sober lab showed that connections which are known to be important for learning motor skills in humans and other mammals are also found in regions of the songbird brain that are required to learn song.
"Thalamostriatal and cerebellothalamic pathways in a songbird, the Bengalese finch"
David A Nicholson, Todd Roberts, Samuel J Sober. on Biorxiv:
"Comparison of machine learning methods applied to birdsong element classification". Proc. of the 15th Python in Science Conf (SciPy 2016)
"Neural networks for segmentation of vocalizations" with Yarden Cohen. PyData NYC 2017. slides on SpeakerDeck
"Support Vector Machines: for the birds?" Curly Braces 2015.