Hello!

My name’s Ben Giacalone, and I’m a CS student at the Rochester Institute of Technology (BS/MS, Spring 2024). I’ve worked primarily with machine learning (PyTorch), web (Vue, Svelte, Django) and gamedev (Godot, Bevy), but I’m always open to learning new stuff. In particular, I have a strong interest in using reinforcement learning to create intelligent systems.

Here are some things I’ve done, some of which are WIPs. Check them out!

Projects

Apps and Demos

DQN Demo

A Deep Q Network demo created for an RIT AI talk.

Visualizing ColBERT Queries

Visualizes ColBERT embeddings and shows what they match to.

Note: Takes a while to load.
Baleen Visualizer

Visualizes the outputs of each Baleen stage.

Mini Research Projects

SmashRL

(WIP) Training a neural network to play Super Smash Bros. Melee with retrieval.

PaintRL

(WIP) Using RL for stroke-based artistic rendering.

Papers

Beneath the [MASK]: An Analysis of Structural Query Tokens in ColBERT

B. Giacalone, G. Paiement, Q. Tucker, and R. Zanibbi

ECIR 2024 | (PDF)

Tools and Libraries

  • RL Template - A standard template I use for RL projects.
  • ViewGD - A Vue-like reactivity framework for Godot 4.