Its been a long while since I last posted, but for good reason! I was busy shifting base from Google’s Hyderabad office to their new location in Sunnyvale. This is my first time in the USA, so there is a lot to take in and process!

When I wrote the original post, I wasn’t planning on writing a follow-up. Certainly not the week after. But what a difference a week can make in a dynamic system like the US stock market.

Here, I give a quick review of the concept of a Martingale. A Martingale is a sequence of random variables satisfying a specific expectation conservation law. If one can identify a Martingale relating to some other sequence of random variables, its use can sometimes make quick work of certain expectation value evaluations.

Ensemble learning methods are widely used nowadays for its predictive performance improvement. Ensemble learning combines multiple predictions (forecasts) from one or multiple methods to overcome accuracy of simple prediction and to avoid possible overfit. In the domain of time series forecasting, we have somehow obstructed situation because of dynamic changes in coming data. However, when a single regression model is used for forecasting, time dependency is not the obstacle, we can tune it at current time of a sliding window. For this reason, in this post, I will describe you two simple ensemble learning methods - Bagging and Random Forest. Bagging will be used with combination of two simple regression trees methods used in the previous post (RPART and CTREE). I will not repeat most of the things mentioned there so check it first if you didn’t make it already: Using regression trees for forecasting double-seasonal time series with trend.

With the abundance of deep learning frameworks available today, it can be difficult to know what to choose for any particular application. Given the contrasting strengths and weaknesses of these frameworks, the ability to work with and switch between more than one is particularly important. Recent Cloudera blogs have shown how examples of applying deep learning on the Cloudera ecosystem using popular frameworks Deeplearning4j, BigDL, and Keras+TensorFlow. This post extends those examples by providing details on performing image recognition with convolutional neural networks using the Apache MXNet (incubating) framework, leveraging Cloudera’s Data Science Workbench for an interactive development environment.

I spent about half of his year so far devoted to my first major research project at Stanford - DeformNet. The gist of the project was to create 3D models of objects based on a single 2D image of the object, by deforming the 3D model of a similar object. It took months to get working properly, but all that work eventually led to a paper - “DeformNet: Free-Form Deformation Network for 3D Shape Reconstruction from a Single Image” - with its very own Arxiv post and project page. But, it also led to much more. It led to horrific, bizzare, and delightful failures.

I can actually look up how long I have by logging into my Coinbase account, looking at the history of the Bitcoin wallet, and seeing this transaction I got back in 2012 after signing up for Coinbase. Bitcoin was trading at about $6.50 per. If I still had that 0.1 BTC, that’d be worth over $500 at the time of this writing. In case people are wondering, I ended up selling that when a Bitcoin was worth $2000. So I only made $200 out of it rather than the $550 now. Should have held on.

Introduction

Let’s directly dive in. The thing here is to use Tensorboard to plot your PyTorch trainings. For this, I use TensorboardX which is a nice interface communicating Tensorboard avoiding Tensorflow dependencies.

** Mon 16 October 2017