Week 13 - Markov

            This week we worked on a Markov project. It was fun to utilize things like punctuation detection to work on word associations to generate random sentences. Both of my group members did well on their Markov project this week. We all had similar approaches with different twists. It was interesting to see how many different imports can be implemented to get the same results within our methods.

I like how the lecture briefly tied in the similarities in this sort of thing to systems like ChatGPT that utilize similar principles on a way larger scale. The Markov project tracks single-word associations, but advanced models like this handle multiple tokens and use neural networks to predict the next word or token. But the lecture also mentioned these models often require extensive training data and sophisticated algorithms to achieve high accuracy. Something else that was briefly touched on was the ethical considerations of this data that’s trained by workers in less-than-ideal conditions.

Now for the reviews. My approach hasn’t changed much since last time. I started by glancing over the UML and overview document. Then I dove into the easy code. I wrote constants, fields, getters and setters, using autocomplete when I could. Then I worked on the logic within the special methods. I had a pdf of the overview handy to make sure I adhered to the specifications the methods needed. I ran tests in between methods to see if they passed. And lastly, I went back to check if any code could be optimized and to comment where I thought it was necessary to make proper documentation. One of my group member's approach was to read the prompt carefully. From the prompt, he identified the necessary fields and then started writing the methods in the order they were needed. He began with reading the file, followed by reading the line, processing the word, adding the word, etc. He found the entire project to be straightforward. He acknowledged that his commenting was a bit off as he did this early, but he has since learned more about Javadoc style formatting. I think I’m starting to find a rhythm, so my strategy won’t be changing much. With project 02 in sight, I'm looking forward to exploring new strategies and seeing how I can tackle the challenge independently.