Towards Building Noise-Ignoring Deep Neural Networks via Neural Architecture Search
In this small late-night project I sought to automatically discover neural architectures incapable of fitting to random noise, following an interesting problem from Twitter . Twitter is great as a social platform for academic work!
If we have a neural architecture search system that can seek to minimize (or maximize) a given loss function, we can employ such a system to discover neural architectures incapable of fitting to random noise by altering the loss as shown in Figure 1 below.
A New Upper Bound for the Moving Sofa Problem
The moving sofa problem is a fairly well-known open problem that essentially asks: “What is the largest area of a shape that can pass through a unit width corridor with an L shape?”
Fairly complex sofa constructions have been offered for this problem. Notably, Joseph Gerver presented a sofa with an area of 2.2195 , and Philip Gibbs computationally found a sofa with a seemingly identical shape . However, no rigorous proof has been offered for the optimality of this sofa shape to this day.
Ubuntu 16.04 + GTX 1080 + CUDA 8.0
I have recently acquired a GTX 1080. A few friends have told me that they have been having trouble getting everything working for a standard machine learning setup, so I have decided to document the steps I have taken.
Boot from a UEFI USB. Add “nouveau.modeset=0” to the end of the options script (press F6, then Escape, then type), otherwise you will get a blank screen. Install Ubuntu, quite straightforward.