Toward Believable Acting for Autonomous Animated Characters

Last month I had the pleasure of presenting some of my team’s recent research at MIG ’22. It’s our first publication, on a topic I care deeply about: acting for autonomous animated characters. Why do NPCs in video games seem so far behind, in terms of acting ability, compared to their counterparts in animated movies? How might we make their acting more believable? This is one of those hard, fascinating problems that are just starting to become tractable thanks to recent advances in machine learning. I’ll have more to say about it soon, but for now, here’s a short video that explains the first small steps we’ve taken in that direction:

You can find the above video, our paper, and also a recording of our 25-minute talk on our new site for the project: https://believable-acting.github.io/

Branching sand patterns

Here in SargĂ­, Brazil, when it isn’t raining, we get to take a lot of long walks on the beach. One feature we noticed right away were these unusual patterns just on top of the surface: little clusters of wiggly lines made of light sand that contrasted sharply against the dark, damp, compact sand beneath. Some were small and isolated, while others formed dense networks. We wondered out loud: what were these shapes, and where did they come from? Were they the trails of some tiny worm or crustacean? Detritus tossed up from the digging of underground warrens? That was our best guess on the first day. But the shapes were so tiny— barely wider than a few grains of sand— and we never saw any evidence of whatever life we imagined was creating them.

Later in the week, the weather changed, and the shapes changed too. The lines got longer, and they seemed to favor certain directions more than others. In particular, there was a strong breeze blowing up the coast, and the lines were oriented in the direction of the wind. Also there was something vaguely familiar about the way the shapes branched out and meandered, but I couldn’t put my finger on it.