paulobruno / pietree

A visualization tool of Decision Trees using d3

I love simple and fast code. And to pretend AIs enjoy playing video games as much as I do.

Great programmers write code for humans and computers. Neglect any of those and you will write poor code. For one or the other.

I'm a PhD student in Computer Science at GSSI and a Deep Reinforcement Learning researcher. I always like to try novel approaches and find unusual ways to solve a problem, looking beyond the immediate solutions. As such, some of my results are very very awful. Some others are very very exciting. And I think that's beautiful.

Also - and this is an important warning - don't expect me to create dozens of classes with intricate multi-hierarchy objects. My goal is to write clean, simple, and, of course, fast code. However, I don't see performance as the main target, but rather as the result of treating the computer with care.

I'm usually programming in one of these languages

📄 Latest publications and talks

• WorldDynamics.jl: A Julia Package for Developing and Simulating Integrated Assessment Models
• Exploring Large Language Models Capabilities to Explain Decision Trees
• Unveiling the Key Features Influencing Game Agents with Different Levels of Robustness
• Integrated assessment modeling using WorldDynamics.jl
• Deep Reinforcement Learning in foraging simulation

pietree's Issues

TODO checkpoints

Here are the features that were in the schedule and were not finished:

• Load table from json file
• Pop-up menus with helpful information in every node
• Variable node sizes according to the number of examples
• Checkboxes to turn features on/off
• To show correctness percentage of the whole tree