Jan 12Using Rules with a SpaCy TextCat Classifier for NSFW ContentThis is a short tech note (intermediate level) on how to modify a spaCy NLP pipeline to incorporate rules to override — or guarantee — a trained classification model result. Models can be fragile sometimes: they can be gamed by adversarial users or behave unpredictably under certain contexts of use…NLP6 min readNLP6 min read
Jun 26, 2021Dwarf2Text: A Mediocre Data-to-Text Generation Project Leads to Learning Stuff (Part 1)In this article, I will introduce text generation from data tables, using a project I did with game data as a base from which to explore the problem domain, industry, and tools. It’ll take 2 parts — in part 1, I’ll introduce my project with its many flaws, then introduce…Generation18 min readGeneration18 min read
Oct 16, 2018“Plotto”: Generating Truly Offensive Stories Since 1928.Tl;dr: I wrote a lot of code to generate stories from the 1928 writer’s manual “Plotto,” and then I looked at the data and decided that just because I could, it wasn’t a good idea. I’m not releasing that dataset and code. There are 2 primary parts to this: The…Feminism28 min readFeminism28 min read
Jun 3, 2018My AI-Art Talk for ResonateEarlier this year, I was one of the folks invited to speak at Resonate Festival who didn’t end up going, after they dropped out of touch in a sad funding fiasco. …Machine Learning8 min readMachine Learning8 min read
Published in data_emlyon·Mar 5, 2018OpenVis Conference: Data Visualization on the Open Web“OpenVis is massively influential and packed to the gills with top talent.” — David Mimno, Assist. Professor, Cornell Information Science OpenVis Conference is a boutique event for data visualization practice. After 5 years in Boston with Bocoup, our primary sponsor in 2018 is emlyon business school, my employer, where I…Data Science16 min readData Science16 min read
Nov 5, 2017COCO’s Memory Palace: A Strange Fantasia(This is my KIKK Festival 2017 talk written up, from the Tech Track. It was opposite @quasimondo’s talk, so I think it might find a larger audience in text.) This talk springs from my interest in transcoding text and images for artistic purposes. This talk is about photographs and generative…Machine Learning14 min readMachine Learning14 min read
Sep 30, 2017Things I Think Are Awesome 4: Maps and Poetry and PeopleIt’s been a long time since the last newsletter/post. Sometimes it’s hard to focus on the awesome, when your course is killing your students, IT doesn’t think it’s their job to have VGA adapters, & you’ve lost your shit several times in the week; and meanwhile people are living without…AI6 min readAI6 min read
May 1, 2017Things I Think Are Awesome, Ep3: Time, Space, Memory[Cross-posted from my tinyletter list here.] Hi. This episode is a lightly linked set of things that fell all over me recently. There’s time/space/memory and some games and generative poetry in here. At last week’s excellent OpenVis Conf (which I co-chair, truth in advertising), Matt Brehmer from Microsoft gave a…Machine Learning6 min readMachine Learning6 min read
Apr 30, 2017Things I Think Are Awesome Ep 2: Bots, MeteorsHi again — thanks for signing up [originally posted on my tinyletter]! This time has more focus on things that make me smile, because they are the most awesome: a lot from Twitter as usual, and also one travel tip. Via a tweet from @DashingD3, I love the Dark Sky…Bots3 min readBots3 min read
Apr 30, 2017Things I Think Are Awesome-Episode 1.[This was originally published as my first post on my new tinyletter. I’ve decided to cross-post them all here since the lack of feedback on tinyletter makes me a little sad.] Like many of us, I’ve had a hard few months online. I remember when the top news stories in…Artificial Intelligence3 min readArtificial Intelligence3 min read
