Nov/110
A Simple Metric for Turn-Taking in Emergent Communication
I'm pleased to announce the online-first publication of our recent research on human and machine turn-taking:
A Simple Metric for Turn-Taking in Emergent Communication
By Peter A Raffensperger, Russell Y Webb, Philip J Bones and Allan I McInnes
(doi:10.1177/1059712311421831)
Abstract:
Oct/110
The greatest bug of my PhD so far
Today I accidentally coded a bug, fixed it, and then intentionally reintroduced it into my code. It was the greatest bug of my PhD so far.
I wrote a routine that check if all Nash equilibria in a set belonged to a particular class, C. That was a great success and, combined with months of effort before, it gave me the best simulation results I've achieved so far in my PhD. That part makes sense, it works right, and it's not the bug.
May/110
Book review: “A New Kind of Science” by Stephen Wolfram
Wolfram examines the behaviour of simple cellular automata and other similar simple systems including multi-dimensional cellular automata, mobile automata, continuous automata, and network system. Wolfram finds that complex behaviour arises from even simple rules, defying the common intuition that complex behaviours must imply complex governing rules. Wolfram claims to have invented a new science and that his discovery of complex behaviour emerging from simple rules will change the way people approach science.
A trivial summary of the book with over one thousand pages is:
Watch out! Many strings that seem to have high Kolmogorov complexity actually have low complexity because of the great expressiveness of even simple rules. Furthermore, there exist cellular automata rules that have Turing complete computational power, which implies that any kind of computable behaviour is possible – a simple set of rules that have the power to express the most complex deterministic systems.
Jan/110
Improved Secure Password Card
It's interesting how far you can go with a design project. It never seems quite prefect. My initial attempt at solving my password generation and memory problem worked fairly well, but I still wanted to improve it.
The excel file (code_cardv2.xlsx) and results are posted below. The list of minor improvements are as follows:
- Slightly better usage of the letter 'y' (now used as both a vowel and a consonant, but not both).
- The center portion of the card (inside the dotted lines) is shared with an adjacent card so that you can manage shared passwords with someone else. You can destroy the other card and ignore this feature if you don't need it.
- The lower right number is only a single digit so that I can round the corners of the laminated card without cutting a number.
- The grid of locations on the back has been replaced with a spinner. The original grid was fine for choosing starting locations, but I felt it potentially gave away information about how you used the card. While this is more of a theoretical concern it still bothered me. Now you punch or drill a tiny hole in your card and you can spin it by inserting a pencil or pen in the hole. Alternately, you could simply flick the card, but I found this requires more practice than I would have imagined. The spinner has a selection of options so you can choose columns, rows, small numbers, patterns of reading, and directions. The intention of the arrow is to choose a random direction (not as you might first think, to point to the result --- obviously that would be useless.)
- A name for this thing (so people can google it), copyright, and URL (which eventually will point to all the resources and instructions for the card), because I think more people should use it.
Missing features:
- Registration of the front to the back when printing is fairly hard to setup in Excel and probably depends on your printer and margins. Since the cards have to be generated randomly, I'm not sure of an easy way to solve the problem. Perhaps gluing to pages together after printing is the simplest way. I could lay it out so you could fold the paper, but then laminating and cutting works less well. Maybe cut, fold, glue, and then laminate would work. What glues can go through a hot lamination system?
May/100
Paper summary: Interconnected musical networks: Toward a theoretical framework
Weinberg, G. (2005). Interconnected musical networks: Toward a theoretical framework. Computer Music Journal, 29(2):23–39
Weinberg (2005) discusses musical networks, the concept of performance as an interdependent art form. Electronics expands the range of possible interdependencies in musical performance: ‘Although acoustic-interdependent models provide an infrastructure for a variety of approaches for interconnections and interdependencies among players, they do not allow for actual manipulation and control of each other’s explicit musical voices. Only by constructing electronic (or mechanical) communication channels among players can participants take an active role in determining and influencing not only their own musical output but also that of their peers’ (Weinberg, 2005, p. 23).
May/100
Book Summary: “Musicophilia” by Oliver Sacks
Sacks, O. (2007). Musicophilia: Tales of Music and the Brain. Knopf.
Sacks explores the neurology of music and gives examples of patients who have various unusual responses to music. Told in a wonderful, person-focused, story-telling manner, Musicophilia expounds on music as the wonderful back door to our minds. One man gained a sudden appreciation of piano music after being struck by lightning. Some people get seizures from music or hear music in their seizures. Most people experience getting music stuck in their head, but some have more intense musical hallucinations.
Apr/100
ChucK: mind-bending programming language of the day
I've recently started teaching myself a new programming language: ChucK. You can read a paper by the authors that gives a quick outline of some of the salient features of the language:
Wang, G., Cook, P. R. (2003). ChucK: a concurrent, on-the-fly audio programming language. In Proceedings of International Computer Music Conference, pages 219–226
Mar/100
Robotic drumming with machine learning back end
This video is what I'd like to build: Jazari drumming robots controlled by Wiimotes and machine learning.
Of course, I'm in the business of building better and more algorithms for music and other purposes, but this is the kind of hardware platform that I'd be keen to demo algorithms on.
You can read more on the artist's website.
Peter
Jan/100
TED talk on evolving robots
Today I found this TED talk on evolving robots by Hod Lipson at Cornell.
Towards the end of the talk he makes the interesting point that in absence of a particular reward, a heterogeneous population of simulated robots ends up favouring those kinds of robots that can self-replicate.
Peter
Jan/100
The Unreasonable Effectiveness of Mathematics
Today I came across and insightful discussion on a dark-ish corner of the philosophy of science: why does math describe the universe so well?
The Unreasonable Effectiveness of Mathematics by Richard. W. Hamming (of Hamming distance and Hamming window fame). Originally this article appeared in The American Mathematical Monthly Volume 87 Number 2 February 1980.
It's a bit of a long read, but here are some highlights: