Matthias Bachfischer

Matthias Bachfischer

Senior AI Solution Architect | E.ON

Building a personal Twitter Bot using Markov chains

less than 1 minute read

Published:

A short announcements that I have now created my own alter ego in the form of a Twitter bot. It goes by the handle of @Bachfischer_bot and you can follow it on Twitter.

If you want to interact with the bot, you must include #FamousQuotesFromRoding in your tweet and it’ll reply to you (Roding is a small village in Bavaria and also my hometown).

The bot works by using Markov chains, which can generate text that looks superficially good, but is actually quite nonsensical. I trained it on Adam Smith’s book “An Inquiry into the Nature and Causes of the Wealth of Nations”, so don’t be surprised if it engages with you in a discusson on free markets and economic affairs.

At some point in the future, I might consider to improve the bot to work with GPT2 (or another language model) or write up a more detailed blog post on the tech stack that I used to deploy the bot (Google Cloud App Engine).

You May Also Enjoy

Testing the robustness of learned index structures

6 minute read

Published:

The database community has recently seen a massive surge in research to replace traditional database indexes such as B+ Trees with machine learning models (also called “learned index structures”) to facilitate fast data retrieval.

Image Localisation via Feature Matching

3 minute read

Published:

In this blog post, we would like to share some insights on our participation in the COMP90086 Computer Vision Project at the University of Melbourne (Semester 2, 2021).

Twitter Rumour Classification with BERTweet

1 minute read

Published:

In this blog post, I would like to present my submission to the COMP90042 Natural Language Processing Project at the University of Melbourne (Semester 1 2021).