Name: Navid Dianati
Type: User
Company: Broad Institute of Harvard and MIT
Bio: Physics PhD with a range of research interests in applied mathematics, scientific computing, algorithms, network analysis and machine learning.
Location: United States
Navid Dianati's Projects
Text synthesis bot that learns from a given text corpus using an LSTM-based deep neural network and produces new text given a "seed" prompt.
Maximum entropy significance measure/filtering algorithm for bimodal affiliation graphs
Force-directed graph layout using velocity Verlet integration.
Deep learning framework for predicting mechanism of action of test compounds based on L1000 differential gene expression profiles of FDA approved drugs. Full data prep, training, testing, validation and analysis stack.
This script generates pdf files that display all the named colors defined by Gnuplot at various alpha values (requires version >= 4.7)
generate a 3D animation of graphs.
Package for "pruning" weighted complex networks based on the Marginal Likelihood Filter.
Library for the analysis of networks
Simple tool for calculating and visualizing costs associated with a mortgage
Python API for creating browser-based reveal.js presentations.
A somewhat customizeable implementation of a seq2seq model based on a keras.io example
A Latex and Lyx template for doctoral dissertations that conforms to the guidelines of the Rackham Graduate School of the University of Michigan.