Chances with Wolves (31/10/2011 11:15 PM)

3 Jul

There’s a new computer model of a virtual wolf pack that can replicate a real pack’s behaviour using just two rules: each wolf moves towards the prey until it reaches a certain safe distance, then moves away from any other wolves that have reached that distance. Following only these rules, the simulated wolves will close in on “prey” and then encircle it. If the prey turns aside, a wolf maintaining its distance will block the prey’s path, effectively ambushing it.

Over fifty years ago, the amateur film-maker Tony Rose found the best way to get natural rather than stilted self-conscious behaviour in front of the camera from his non-actors was for most of the scenes in which they appeared to be played like games, with him as referee. For a scene in a skiff, he told each of them in turn that the game was to stay in the boat as long as possible while trying to get everyone else into the water.

When making “The Kid”, ninety years ago, Charlie Chaplin spent hours playing with his six-year-old star Jackie Coogan. The boy was a natural mimic. They would rehearse each scene until, recalled Chaplin, “eventually, he was so sure of the mechanics that his emotion came with them. In other words, the mechanics induced the emotion.”

The computer model doesn’t tell us what motivates wolves to adopt this pattern of hunting. The non-actors didn’t actually know what their movie was about. Jackie Coogan was able to convince the audience of his character’s feelings because he was behaving in exactly the same way as someone would who really did have those feelings.

The article in the New Scientist reminded me of the second and third examples. There are times when setting up a game, using rules to constrain the action rather than developing any deep understanding of the sources of the action, can still lead to the required outcome. Maybe it’s something we can bear in mind here when trying to devise more intelligent or apparently responsive systems – they may only need to look like they know what they’re doing.

Scientists rely on the same powers of insight as anyone else who’s got a problem to solve. You can steal from anywhere.

Go to the movie about the factory if you want to see what’s wrong with the system – don’t wait to buy the book about what Toyota did, you’ll be fifty years too late – figure it out for yourself.

Give it another fifty years, though, and computers will be able to fake many of the sorts of insight scientists have now well enough for the difference not to matter. The machines will just look like they’re thinking, but it’ll work anyhow.

So hurry while you’ve still got the chance to play.

Advertisements

2 Responses to “Chances with Wolves (31/10/2011 11:15 PM)”

  1. drew stephenson July 4, 2014 at 8:59 pm #

    reminds me of the comment about sufficiently advanced technology being indistinguishable from science. Does it matter how it works if it does?

    • pabloredux July 4, 2014 at 9:35 pm #

      I’ll have to see how I replied in 2011. I think a lot of people may secretly prefer a sufficiently advanced form of conjuring that’s indistinguishable from magic?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Emma C. Robinson

Researcher in computational methods for medical imaging; affiliated with the (developing) Human Connectome Projects; author of the Multimodal Surface Matching (MSM) software

Troovus

Social uniformity and sameness harass the individual most

Hack Circus Podcast

Behind the scenes of creativity

Chris Glynn

the thinking eye

Naturalistic Philosophy

Teleosemantically speaking

The Captured Thought

~The subjective experience of thinking

Utopia or Dystopia

where past meets future

i heart literati

trivia | brain science | a series of intrigues

Gravity and Levity

A blog about the big ideas in physics, plus a few other things

A HEAD OF DEPARTMENT’S BLOG

I'm maintaining this blog in my time as Head of English, University of Exeter. It's intended to reflect on matters relevant to colleagues and academics elsewhere: research, teaching, national policy, university life, managing our time, and so forth. All views are my own and not necessarily those of my employer, nor my department. If you like a post, please tweet to the world; if you don't please leave a comment to tell me why - Andrew McRae (http://humanities.exeter.ac.uk/english/staff/mcrae/; @McRaeAndrew).

Form Follows Function

All Things Architectural

The Dragon and The Phoenix

Bard's Journey into the Collective Subconscious and How It Shapes Our World

Praxtime by Nathan Taylor

The near future. The next 5-20 years.

George Zarkadakis

The official blog of author George Zarkadakis

Thomas Hurka

Professor of Philosophy, University of Toronto

Sean Enda Power

Professional work: articles, talks, public, interdisciplinary, and academic (includes work from 'Time and Illusion')

A Mind for Madness

Musings on art, philosophy, mathematics, and physics

Deep Safety

Victoria's musings on AI safety, machine learning, and rationality.

itakoyak

Living in Japan and posting some AI research thoughts.

Si Alhir (Sinan Si Alhir)

© 2016 Si Alhir (Sinan Si Alhir) — All Rights Reserved.

The Disorder Of Things

For the Relentless Criticism of All Existing Conditions Since 2010

Abeba Birhane

A little bit of Everything. Cognitive Science. Ethiopia. Ireland.

Writing my own user manual

Sergio Graziosi's Blog

Theory, Evolution, and Games Group

The EGG studies theoretical computer science, evolution, and game theory.

Filosofia: estudo e ensino

por Marcelo Fischborn

Charlie Alfred's Weblog

Thoughts on Software Architecture, Value Modeling, and Communicating about Complexity

Refugee Archives @ UEL

news and developments

The knowledge relation

Notes from a radical empiricist

Thom Scott-Phillips

Senior Research Scientist, Central European University, Budapest

Curious About Consiousness?

Me too! All thoughts belong to Heather Iriye, a PhD student in Psychology at the University of Sussex passionate about public engagement with science.

A Shared Madness

A FILM BLOG TO KEEP YOU SANE!

Nucleus Ambiguous

mind of science / science of mind

Samir Chopra

Refusing to Stick to the Subject

Sectioned

A blog about mental health & mental healthcare

Mind Hacks

Neuroscience and psychology news and views.

Bulldozer00's Blog

Exploring And Discovering

Weskaggs

Evolution: it's nothing personal

slehar

A topnotch WordPress.com site

Conscience and Consciousness

Academic Philosophy for a General Audience

Rolf Zwaan

Evolution: it's nothing personal

PhyloBotanist

Evolution: it's nothing personal

Wiring the Brain

Evolution: it's nothing personal

P-world and R-world

A solopsistic musing. Not terribly accessible.

The Elusive Self

Comment on the science of the self

%d bloggers like this: