Paper Reviews
Paper Review: An Integration of Deep Learning and Neuroscience, Part 3

Paper Review: An Integration of Deep Learning and Neuroscience, Part 3

In our previous post we talked about how the brain might optimize cost functions. Now we’ll explore how cost functions may be generated, represented, and change over time in the brain. Marblestone et al outline several ways that cost functions could be generated.  In particular, they talk about specialized circuitry for comparing the predicted output...
Paper Review: An Integration of Deep Learning and Neuroscience, Part 2

Paper Review: An Integration of Deep Learning and Neuroscience, Part 2

One of the key sticking points in discussions comparing machine learning and the brain is how the notion of “learning” differs between computational and biological systems. In section 2 of their paper, Marblestone et al. grapple with this issue in detail. For our introduction post on this paper, go here. Deep neural networks are trained...
Paper Review: An Integration of Deep Learning and Neuroscience, Part 1

Paper Review: An Integration of Deep Learning and Neuroscience, Part 1

Nicole: More and more I see that people are very concerned with the biological plausibility of neural networks. I think this comes from the fact that we as machine learners are finally achieving human-level performance on some tasks. It has renewed faith in the idea that the best way to “solve” intelligence is to copy...
Paper review: The brain adapts to dishonesty

Paper review: The brain adapts to dishonesty

Does dishonesty escalate over time? Garrett et al seek to answer this question and substantiate something we all have observed anecdotally: that dishonesty can be a “slippery slope.” The idea is that if you are dishonest once, it is easier to be dishonest again, and this effect can accumulate over time. This paper generated some...
ML Ethics and the "Criminal Detector"

ML Ethics and the “Criminal Detector”

Confession: I roll my eyes whenever someone says the word ‘singularity.’ As awesome frightening as it would be to have computers take over the world, the fear of that potential future distracts from the more pressing ethical concerns surrounding machine learning and AI today. Edward Snowden alerted everyone in the US to the possibility that...
Paper Review: Fixing the stimulus­ as fixed ­effect fallacy in task fMRI

Paper Review: Fixing the stimulus­ as fixed ­effect fallacy in task fMRI

Fixing the stimulus­-as-­fixed-­effect fallacy in task fMRI Jacob Westfall, Thomas E. Nichols, and  Tal Yarkoni http://biorxiv.org/content/biorxiv/early/2016/09/25/077131.full.pdf In traditional fMRI experiments it’s typical to show subjects a limited set of stimuli (e.g. “tasty tomato”), but generalize the results to larger classes from which the stimuli were drawn (e.g. adjectives and nouns). According to Westfall, Nichols and...