Talking Games and Graphs at NODES 2022

In November 2022, I had the opportunity to share my research in long-term memory for non-player characters in games. Neo4j’s yearly international conference, NODES 2022, attracts graph database experts from around the globe to share their work and network with like minded thinkers in the area of graph databases, graph machine learning, and applications of graph database solutions to a massive array of use cases.

“Everything is a graph” is an oft repeated mantra amongst Neo4j enthusiasts. This mantra resonates for me because I study relationships of all types. I was recently asked to provide an artist’s statement for a podcast I was invited to participate in with my friends @minbid (https://www.instagram.com/minbid/) which allowed me to share my artistic perspective on the work I do. This is what I shared:

Relationships are everything and everywhere. I am fascinated by relationships of all kinds regardless of what objects are involved or whether the relationships take a corporeal form or exist in the ether of our minds. I build objects that exist in the real world and the virtual. Sometimes I merge the two in the same space and time. I use tools that are made of code, light, sound, plastic, paint, tape, or any other form that allows me to connect with you.

For me, graph databases provide an opportunity to model and experiment with relationships in data in ways that traditional relational databases can’t (or if they can, it is challenging to generate the schema/query cycle that makes iterative adaptations difficult). In graph databases, the killer app is in the relationships between nodes. I tend to think of nodes as the nouns and adjectives, while the relationships are verbs and adverbs (yes, we can add nuances to the relationships). This parallel is just a starting place, so I try not to get too hung up on the details of the direct correlation between language structure and graph database structure. As well, in my work with games and 3D, I find that I can visualize my data in a 3D graph or manifold.

With NODES 2022, I participated in two sessions. First, I was invited to a panel discussion to discuss Neo4j and the impact it has had on my work. Second, I gave a 30 minutes presentation of my doctoral research demonstrating the ways that I use a custom created plugin to connect Neo4j to Unreal Engine. This plugin was designed and engineered by myself and my collaborator, Michael Wahba, and uses a JSON/REST based connection that allows continuous independent queries to be sent from the real-time gamespace and received from the database. With it we can track all manner of data in game and use synchronous and asynchronous machine learning algorithms to empower the agents in the simulation (the NPCs in games) to learn from and adapt to their environment on the fly.

My learning continues! At NODES, I learned about the work of Dr. Kirell Benzi who describes himself as a data artist. This resonates for the artist network in me. The key insight from his data artist manifesto is the idea that visualizations tend to impose a chart model on the data, where data artistry allows the shape of the data to reveal itself in emergent ways. <<cue gong resonance sound FX>> As well, I was fascinated by a talk that explored hypergraphs in the context of sports teams and referenced an amazing talk from a few years ago that used hypergraphs in Neo4j to model the Marvel Comics Universe.

NODES 2022 was an excellent conference where I connected with many researchers, developers, data scientists, and graph database enthusiasts. The feedback from my presentation was inspiring and insightful. If you have the opportunity, I highly recommend looking at the NODES conference YouTube stream. You might just get inspired yourself! You can find a full list of NODES 2022 talks on YouTube.

PS - Delighted to share that I am included in the “20 NODES 2022 Talks You Don’t Want to Miss”