Personal tools
You are here: Home Events ANC Workshop Miltos Allamanis/Sander Keemink Chair Daniel Trejo Banos

ANC Workshop Miltos Allamanis/Sander Keemink Chair Daniel Trejo Banos

— filed under:

  • ANC Workshop Talk
When Feb 17, 2015
from 11:00 AM to 12:00 PM
Where Room IF 4.31/4.33
Add event to calendar vCal



Title:   Neural Logbilinear Models for Learning Naming Conventions in Source Code

Every programmer has a characteristic style, ranging from preferences about identifier naming to preferences about object relationships and design patterns. Coding conventions define a consistent syntactic style, fostering readability and hence maintainability. When collaborating, programmers strive to obey a project’s coding conventions. However, one third of reviews of changes contain feedback about coding conventions, indicating that programmers do not always follow them and that project members care deeply about adherence. Unfortunately, programmers are often unaware of coding conventions because inferring them requires a global view, one that aggregates the many local decisions programmers make and identifies emergent consensus on style. We present a neural log-bilinear model that is able to learn about naming conventions of variables and methods and can suggest stylistically consistent names. We also show that the logbilinear model is able to produce semantically consistent embeddings, as in word2vec.

Title: A story of minimum curvature energy and visual perception
In 1691 James Bernoulli wondered what shape a bendable rod of negligible weight would make if a weight was hanged from one end. This kickstarted the theory of elastica, which describes the lines of minimum curvature energy between some start and end orientations and locations. A straight line has no curvature, and zero energy. As one bends a line in various ways, the curvature increases and so does its energy.

In my talk I will briefly discuss the theory of elastica, and then link it to cell interactions in the visual cortex. The resulting model portrays many aspects of cell responses and perception, in particular explaining some counter-intuitive illusions. I thus propose elastica as an underlying principle of the visual cortex.