-
ASCII - Unicode • Chaos
So, where to begin ? I recently (at the beginning of this year) designed a multi-lingual website as a chapter of discovery surrounding my AlphaBet research project – (several years working at Google as a software engineer designing transliteration tools and algorithmic improvements to allow people to use various keyboards to map between languages) – and was surprised to discover how certain web hosts continue to fumble Unicode.
-
Things I Wish People Told Me
Things I wish people told me about working with computers
-
(Art)ificial Art
Some experiments in (Art)ificial Art : generative machine learning algorithms.
-
Exploring ChatGPT in Arabic
An exploration of what ChatGPT can handle with Arabic language.
-
Original Internet
The Silk Road :
-
Leaving Google
I recently left Google after five and a half years working as a software engineer.
-
VGG2Vec: Extracting Image Representations from Neural Networks
In this post I will introduce VGG2Vec, a method to create a vector representation of images using neural networks. By extracting the vector representation of an image from a trained neural network, you can take advantage of what the neural network has learned about the picture. We’ll visualize our representations to show that they encode information about different artists’ styles as well as subject matter.
-
Generating images in different styles with neural networks
Neural networks have been getting deeper and deeper, leading to a range of techniques to try to understand how they work. I wrote about some experiments with one of these, Deep Dream; this post will introduce another called Neural Style, introduced by Leon Gatys, Alexander Ecker, and Matthias Bethge of Bethge Lab.
-
"What Can Be Done?": Sonya's Suffering and the Power of the Word
I got a bit of a blast from the past this week when I received a copy of The Birch Spring 2015, which contains an essay I wrote in the fall of 2012.
-
Seeing through Google's eyes: neural nets and computer vision
Many of the latest advances in image and audio processing (Facebook’s DeepFace, for example, or Google’s voice search) have come from developments in artificial neural networks, or “deep learning.” Neural networks are a type of statistical machine learning model inspired by the workings of the human brain. Like the human brain, their functionings are a little opaque, and we have little idea why they work so well. By exploring Google’s Deep Dream visualization approach, this post will introduce you to neural networks and some of the cool things they can do.
-
Reflections on SemEval2015
As a part of the Natural Language Processing class I took with Rich Wicentowski last fall, teams of students competed to build machine-learning models that could accurately predict the sentiment of Twitter data for SemEval 2015, the International Workshop on Semantic Evaluation. Given that I spent a summer building commercial machine classifiers, I’ve had some time to reflect on the first machine learning model I built.
-
Runkeeper and the collection of personal health data
We stand at the edge of a dive into the Internet of Things, where all devices speak the language of data. One of the most interesting applications of this technology is tracking our own bodies to learn what and how we’re doing.
-
From zero to Hexo: a quick tutorial
Hexo bills itself as a “A fast, simple & powerful blog framework”, but I was blown away by how quickly I was able to get Hexo up and running. The main issue I had was finding the information I needed all in one place (this has been done before, with varying success).
-
Screen in two minutes
I wrote up a quick how-to that I wished I had found when I was first learning screen, instead of having to navigate the vaguely terrifying documentation (“Each virtual terminal provides the functions of a DEC VT100 terminal and, in addition, several control functions from the ISO 6429 (ECMA 48, ANSI X3.64) and ISO 2022 standards (e.g. insert/delete line and support for multiple character sets)”).
-
Swarthmore Course Review Launch
Long story short: a group of students got together and built something pretty awesome, the Swarthmore Course Review, or SwatCoRe for short.
-
ReportIt wins overall prize at TriCo Hackathon
I’m pretty thrilled to share that my team won the overall prize at the 2015 TriCo Hackathon. It was my first hackathon, so I wasn’t very sure what to expect, but my friend Sedinam, who was also one of the two other people on my team, talked me into it.