- Primer on Information Theory and Hypothesis Testing
- This report explains how many common hypothesis tests can be thought of from an information theoretic perspective. The material comes mostly from Cover and Thomas’s Information Theory
- Markov Chain Mixing and TVD
- This report examines how the mixing of Markov Chains can be quantified using TVD
- Time-Frequency Analysis using the Hilbert Huang Transform
- This report explains the Hilbert Huang Transform and its statistical properties.
- Taxi Prediction and Optimal Policy
- This report explored whether it was learn an an optimal policy for taxi drivers-where to go to earn the most fares. This first involved forecast demand for NYC Taxis across Manhattan using ARIMA and LSTMs. Our method then took into account demands in adjacent boroughs and their dynamics over the year and how this increased or decreased traffic flow. And finally this was fed into a policy to how much a taxi driver could make in a day based on this.
- Beyond the Gaussian: Graphical Models
- This covers the development of a crucial tool in nonparametric graphical models, the nonparanormal graphical models. It also includes an augmentation of the method using Binary Expansion Test statistics.
- Holo-Hilbert Huang Transform (Part 1) (Part2)
- This illustrates how the Holo-Hilbert Huang Transform, a multidimensional Extension of the popular Empirical Mode Decomposition can be used in situations where the Fourier and or Wavelet Transforms are inappropriate.
- Stochastic Gradient Descent and its Variants
- This its an overview of SGD and the popular variants that have been used to speed up Machine Learning Optimizers.