I’ve been making my way through the recently released Deep Learning textbook (which is absolutely excellent), and I came upon the section on Universal Approximation Properties. The Universal Approximation Theorem (UAT) essentially proves that neural networks are capable of approximating any continuous function (subject to some constraints and with upper bounds on compute).

There are a variety of tools that can be employed to solve cryptograms, just like there are lots of techniques that can be used to solve puzzles like Sudoku. These include: Algebraic tehniques, number theoory, factorization, digital roots, elimination, carries, divisibility rules, modular arithmetic, the Chinese Remainder Theorem, and even simple trial and error.

The AUC score is a popular summary statistic that is often used to communicate the performance of a classifier. However, we illustrate here that this score depends not only on the quality of the model in question, but also on the difficulty of the test set considered: If samples are added to a test set that are easily classified, the AUC will go up — even if the model studied has not improved. In general, this behavior implies that isolated, single AUC scores cannot be used to meaningfully qualify a model’s performance. Instead, the AUC should be considered a score that is primarily useful for comparing and ranking multiple models — each at a common test set difficulty.

TL;DR: there’s a Shiny app too.

TLDR: In this blogpost, we’re going to train a neural network that is fully encrypted during training (trained on unencrypted data). The result will be a neural network with two beneficial properties. First, the neural network’s intelligence is protected from those who might want to steal it, allowing valuable AIs to be trained in insecure environments without risking theft of their intelligence. Secondly, the network can only make encrypted predictions (which presumably have no impact on the outside world because the outside world cannot understand the predictions without a secret key). This creates a valuable power imbalance between a user and a superintelligence. If the AI is homomorphically encrypted, then from it’s perspective, the entire outside world is also homomorphically encrypted. A human controls the secret key and has the option to either unlock the AI itself (releasing it on the world) or just individual predictions the AI makes (seems safer).

It looks that there is a dominant cluster scattered all over the space: this is mainly due to the general category of news.

Our work titled Online Human-Bot Interactions: Detection, Estimation, and Characterization has been accepted for publication at the prestigious International AAAI Conference on Web and Social Media (ICWSM 2017) to be held in Montreal, Canada in May 2017!

Download the NB: https://github.com/twiecki/WhileMyMCMCGentlySamples/blob/master/content/downloads/notebooks/random_walk_deep_net.ipynb

As a University professor, when setting data orientated projects to Computer Science undergraduates, I used to find it difficult to get students to interact properly with the data. Students tended to write programs to process the data, produce a couple of plots, but fail to develop any understanding of what useful information was in the data, and how best to extract it.

This might look a little funky for those used to Bayesian parameter estimation for univariate systems - estimating $P(\text{heads})$ given the results of 15 coinflips, for example.