Author Archives: underwhelmingforce

Adventures in Reinforcement Learning: Trials and Errors

Posted on May 8, 2018 by underwhelmingforce

Over the last few weeks, I’ve been working on learning some basic reinforcement learning models. All my code is available here (warning: basically totally undocumented and uses code from a few sources. It’s not the prettiest). The Environment The environment I … Continue reading →

Posted in Reinforcement Learning | Tagged , , , , , , , , , , , , | Leave a comment

Unexpected results in ML & project updates

Posted on April 24, 2018 by underwhelmingforce

There was a recent paper that caught my eye that I think is worth sharing with anyone interested in the themes of my blog: The Surprising Creativity of Digital Evolution: A Collection of Anecdotes from the Evolutionary Computation and Artificial … Continue reading →

Posted in Uncategorized | Tagged , , , , , | Leave a comment

Sentiment Translator

Posted on March 14, 2018 by underwhelmingforce

Introduction A very common problem in natural language processing is that of sentiment analysis, or rather, attempting to analytically determine if a given input has an overall positive or negative tone. For example, a positive string might be “I really enjoyed this movie!” whereas … Continue reading →

Posted in nlp | Tagged , , , , , | Leave a comment

Automated topic extraction with Latent Dirichlet Allocation

Posted on January 8, 2018 by underwhelmingforce

During last semester, I became aware of a really interesting NLP algorithm called LDA, short for “Latent Dirichlet Allocation.” The goal of the algorithm is to create a set of “topics” which represent a specific human-interpretable concept The core assumption … Continue reading →

Posted in LDA | Leave a comment

Making art with neural style and fractals

Posted on January 7, 2018 by underwhelmingforce

I recently attempted to see if I could create art with Neural Style using only photos I’ve taken and fractal flames I’ve created with Fractorium and Apophysis. I must say, I’m very happy with the results! I generated the outputs … Continue reading →

Posted in neural style | Leave a comment

Steam Games

Posted on August 5, 2017 by underwhelmingforce

While I work on my next big project, I decided to generate some random steam game names. All of these are games that don’t actually exist: Happy Panic Unraveled Land Mad Sharkness The Heart’s Medicine formic innocence Heroes Over Europe … Continue reading →

Posted in Uncategorized | Leave a comment

Espresso is the marshmallow of coffee (fun with Word2Vec)

Posted on May 30, 2017 by underwhelmingforce

In exploring the recipe dataset, I decided to have some fun with Word2vec, an algorithm originally created by Google. For the layperson, this algorithm works by looking at the context in which a given word appears and learns vectors to … Continue reading →

Posted in word2vec | Tagged | Leave a comment

Chicken soup (for the robot soul?)

Posted on May 24, 2017 by underwhelmingforce

So I did a little more training with the chef-rnn at a lower learning rate to fine-tune it a bit and got some shockingly good results, with some weird quirks. I really just couldn’t resist posting more recipes. These are … Continue reading →

Posted in char-rnn | Tagged | Leave a comment

CycleGAN and Chef-RNN update

Posted on May 22, 2017 by underwhelmingforce

There has been a development in image generation that I find absolutely fascinating. It’s based off Generative Adversarial Networks, which are a very powerful and promising model for image generation. A recent modification to this architecture, the Conditional GAN has been making some … Continue reading →

Posted in char-rnn, CycleGAN | Tagged , , , , , , | 2 Comments

Double Jeopardy

Posted on February 20, 2017 by underwhelmingforce

Because I was gone for most of the weekend, (and because I didn’t want that awful jokebot being the first thing people see of this blog) I retrained the jeopardy network to some amusing results. Take a gander: THE SPORTING … Continue reading →

Posted in Uncategorized | Leave a comment