bonkerfield | opus.stedden

machine learning

Should an LLM write my blog now?

Should an LLM write my blog now? — a.ttent.io/n, a new baby, and what gets lost when you skip the middle

Story2Hallucination: converting stories to deep learning GAN hallucinated animations

Story2Hallucination: converting stories to deep learning GAN hallucinated animations — Using OpenAI CLIP to generate dynamically evolving images as the text of a story is fed in

a.ttent.io/n

a.ttent.io/n — "Out early for drugs and octopus," Ezzy texted. She was sitting on the hallway floor of the Institute's distribution center, head leaning back against the wall. Three seconds after 6 AM she started tapping impatiently ...

analyzing a.ttent.io/n

analyzing a.ttent.io/n — Why do written works of fiction ever have to be finished? What if the story's growth was part of the story itself? My first real work of fiction is finished. But I wonder if my ...

A GPT2+BERT reddit reply bot in 20 minutes

A GPT2+BERT reddit reply bot in 20 minutes — There's a new subreddit called talkwithgpt2bots that's intended to let people interact with AI based comment bots. I got my tupperware-party bot working again on there, but it would be awesome to have a bunch ...

How to build a convincing reddit personality with GPT2 and BERT — Last month, I experimented with building a reddit comment bot that generated natural language replies by combining two pre-trained deep learning models: GPT-2 and BERT. I wrote another post on the motivation and background, but ...

Combining GPT-2 and BERT to make a fake person

Combining GPT-2 and BERT to make a fake person — GPT-2 is a deep learning model that is able to generate astonishingly coherent English text. It was released last year, and everyone’s mind was blown into histrionic hyperbole, including mine. Its creators at OpenAI were ...

Text Mining and Natural Language Processing on Health Forums

Text Mining and Natural Language Processing on Health Forums — As part of the Insight Health Data Science Fellowship, I just got to spend the last 3 weeks working on a pretty fun project applying natural language processing to medical health forums. The goal is ...