Simon Dobson (Posts about complex systems)https://simondobson.org/enContents © 2020 <a href="mailto:simon.dobson@computer.org">Simon Dobson</a> Fri, 27 Nov 2020 17:36:16 GMTNikola (getnikola.com)http://blogs.law.harvard.edu/tech/rss- How computer science can help keep you healthyhttps://simondobson.org/blog/2015/07/01/healthy/Simon Dobson<div><p></p><p>Well, it has to be good for <em>something</em>...</p>
<!--more-->
<p>People sometimes aren't aware just how much computers influence their lives. They've used the internet and mobile phones, seen computer-generated imagery in cinemas, and perhaps realised how much date is being sensed around them. But there are enormous applications for computers in science, arts, and medicine.</p>
<p>Earlier today I did an introductory lecture on using computers to study disease epidemics:
</p><blockquote>Computational epidemiology is the use of mathematical and computational techniques to model how diseases spread. This is important for answering a number of questions. How infectious are different diseases? Why are different populations affected differently? How do different treatment regimes work? Is quarantine effective? We can address these sorts of questions using a range of different techniques, ranging from differential equations (calculus) for simple cases through to complex networks and high-performance simulation for complex case — and possibly even modelling real diseases in real-world geographies in real time.
<p>This lecture is an interactive introduction to these ideas. We’ll explore how diseases spread; conduct an experiment where we infect each other (kind of); and then see how different aspects of computer science help us to explore diseases and their treatment.</p></blockquote>
The slides and other material are available <a href="http://www.simondobson.org/teaching/epidemiology/">here</a>. I've included the slides, and an animation of a simulated epidemic running through a population of people. I've also included an IPython notebook describing some of the mathematics needed and containing all the code I used to generate the graphs and animation from the talk, which might be handy for anyone wanting to explore this area more thoroughly.</div>Blogcomplex systemsepidemic spreadinghttps://simondobson.org/blog/2015/07/01/healthy/Wed, 01 Jul 2015 10:56:05 GMT
- Complex networks, complex processeshttps://simondobson.org/blog/2015/02/13/book/Simon Dobson<div><p></p><p>I'm writing a book on my sabbatical. Or trying to, anyway. So I thought I'd publicise the fact so people can hassle me to keep at it.</p>
<!--more-->
<p>I've been working on complex systems for a couple of years, especially on complex networks: things like the way people move through a road and rail network, or how diseases spread through social networks. It's a bit of a change from my previous work on sensor data interpretation, although not as much as you might think: I'm wondering whether we could combine sensing and simulation, to use sensors to confirm predictions or to drive and condition further simulations.</p>
<p>Getting into this area has been -- and is -- a head-wreck. It's both highly mathematical and highly computational. I understand the computing; the maths, not so much. Many computer scientists would have the same reaction, but conversely, so would many mathematicians: the maths would be familiar, the computing a challenge. So effectively in order to make progress you have to climb two learning curves simultaneously: some unusual and challenging mathematics about stochastic processes, simulated using cluster or cloud computing which poses a lot of challenges even for someone used to programming.</p>
<p>This is made harder by the research literature, though, which tends towards sparse mathematical descriptions, which is frustrating at two levels: the computing is probably interesting (to people like me), and it's hard to re-create the results when the computational approach underlying the graphs and results is unclear.</p>
<p>So with this in mind, and because I've never done it before, I've decided to write a textbook: <em>Complex networks, complex processes</em>. (No, I'm not very imaginative when it comes to titles...) The idea is to link the maths to the code, providing everything a research would need to get started with the maths and the computing. Since this is likely to be a book with, shall we say, <em>limited circulation</em>, I've decided not to bother with a publisher and instead make it completely open. You can look at the current state on the web <a title="Book home page" href="http://www.simondobson.org/research/complex-networks-complex-processes/" target="_blank">here</a>, download the sources, copy and run the code, or anything needed to get started.</p>
<p>It's a work in progress and it's not very usual to advertise books before they're in a fit state to be read, but I suppose that's just a part of open science: make the process visible, warts and all. It also means I'll hopefully get comments and encouragement to keep at it when it starts to fall by the wayside of other things I have to do. The goal is to get the majority done while I'm on research leave (until September), and comments on style, content, and progress will be most welcome.</p></div>Blogbookcomplex systemsnetwork sciencehttps://simondobson.org/blog/2015/02/13/book/Fri, 13 Feb 2015 11:54:41 GMT