Name: William (Bill) Coleman
Type: User
Company: iCIMS
Bio: Data Scientist at iCIMS, where I work primarily on NLP problems. Other focus areas include audio (my PhD domain), bias/fairness and machine translation.
Twitter: _billcoleman
Location: Dublin, Ireland
William (Bill) Coleman's Projects
Active Learning code for an experiment to assess methods for minimising the manual effort required to label audio instances with hierarchical labels.
Config files for my GitHub profile.
Python files written for Experiment 4 which investigated ML performance when predicting auditory hierarchy using CNN and SVM algorithms.
This repository contains EmoBank, a large-scale text corpus manually annotated with emotion according to the psychological Valence-Arousal-Dominance scheme.
Code for a MVP demo IoT device. We used this to send a text message to our future angel investor telling him the current temperature and atmospheric pressure from MEMS sensors in the device. We closed the deal.
Intermediate Natural Language Processing: Real World Applications of Word Embeddings
Online labelling environment to crowdsource labels for auditory stimuli.
TensorFlow implementation of "Multimodal Speech Emotion Recognition using Audio and Text," IEEE SLT-18
Web application to record speech for an open data set
A plugin for recording/exporting the output of Web Audio API nodes
sonicPainter was my masters thesis project. It is a re-imagining of the traditional piano roll as a visual score space based on Xenakis' UPIC. Fundamentally, you draw and manipulate shapes rather than playing fixed notes.
Short R code snippets doing visualisations for the subjective ratings of auditory objects gathered in Experiment 1.
An exploration of the SVHN problem using Keras and a CNN.
Code repository for Think Bayes.
Text and supporting code for Think Stats, 2nd Edition
Repo for Dublin.AI personal project looking at developing a voice biometric monitoring solution for mental health applications.