Strange Loop

A Robot Poet Goes for a Walk in the Park

Designing complex, dynamic systems that can produce interesting and aesthetically pleasing art is a very hard problem to solve, even when you're just talking about something as focused as a Twitter bot. What happens when you try to make procgen art that doesn't just exist on the Internet, but actually lives in and interacts with the real world?

Computational Flâneur is a generative site-specific poetry walk that came out of my research at the MIT Media Lab. As you wander through the Fort Mason Park in San Francisco, a poetry robot makes up and reads you poems based on where you walk. Walk by the bay, and you'll hear poems of the sea; walk by the cannons and you'll hear poems of war.

This talk will explore the design and development of Computational Flâneur. On a technical level, we'll discuss the neural network model that powers the piece's poetry generator, including a brief overview of common text generation tools and how this exact structure was chosen to create a specific "bot-like" aesthetic. We'll also discuss how it's seeded by various forms of smartphone sensor data like location and weather, what sorts of prototyping processes were needed to fine-tune it, and how to conceptually approach this sort of real-world procgen differently from normal procedural generation problems.

Em Lazer-Walker

Em Lazer-Walker


Em is a Toronto-based artist/engineer who makes interactive art, experimental games, and software tools. Most of her work focuses on using nontraditional interfaces to reframe everyday objects and spaces as playful experiences. She’s built projects as far-flung as a site-specific generative poetry walk, a game played on 19th century telegraph hardware, and a commercial board game that uses Amazon Alexa. Currently, she works as a Cloud Advocate for Microsoft Azure.