Computational Mechanics Group Meetings: 2008

Spring Meeting Time -> Wed noon - 1pm

January

16 Jan.

Start-of-quarter organizational meeting

Meetings for this Winter quarter are scheduled for Wednesdays, 1:00 until 2:00.

Ideas for Future Meetings:

• Continue with discussion of Measures of Complexity and create a reference document containing definitions, illustrative examples, and references.
• Develop a FAQ about computational mechanics and its relationship to physics. For example, how might one respond to the question "Why bother with an e-machine when I can already take the Fourier Transform of my data?".
• Develop a series of example processes/machines that attempt to isolate and illustrate particular properties. For example, a machine that has excess entropy but no entropy rate. Or a nondeterministic machine that has a surprising compact deterministic form.
• Tutorial on the basics of the CMPy package (Ellison)

Historical Readings:

• suggestions?

Speakers:

• Raissa (JPC)
• Sean Whalen on research
• Spencer Matthews on research
• Chris Ellison on research/practice oral exam
• Benny Brown on research/practice oral exam
• John Mahoney on basic ideas in quantum information

23 Jan.

Continue with discussion of Measures of Complexity.

Come prepared to describe your favorite measure. If you can, send a TeX entry to John for your measure to be included in the reference document.

Sean gave an example of Lempel-Ziv complexity.

Chris gave intros to several `Difficulty of Creation' measures.

Discussion of `statistical' vs. `deterministic' complexity types. Which measures have to do with the class of data, and which have to do with the particular string observed.

30 Jan.

Computational Mechanics Defense FAQ

As a physics student, I am often asked why what I study is considered physics and also why it is considered worthwhile.

Let's each bring some questions to the table, such as `Why would I ever want to use an emachine instead of the Schrodinger equation?' and see if we can collectively come up with some good answers to these valid questions.

February

6 Feb.

Stochastic Resonance

We will discuss a short paper on stochastic resonance.

13 Feb.

Carl Boettiger will present some of his research and some background.

I will give an overview of my recent work in ecology and evolution. Specifically, we'll discuss a relatively new and controversial approach to evolution called adaptive dynamics. We'll see branching of traits, evolutionary suicide, bifurcations, hysteresis, and evolutionary oscillators. We'll also look at some of my own work on the repeatability of evolutionary trajectories.

Here is a 3 page intro to adaptive dynamics.

20 Feb.

Kristin Lui will talk about Q-learning.

27 Feb.

Chris Ellison presents some of his work.

This Wednesday, I will go over recent work involving the computation of excess entropy via type classes. In the 45 min or so, I will introduce the basic definitions/quantities and go through, in detail, an illustrative example. Additional thoughts will be provided. Comments (especially from John) and are encouraged and expected. A goal of this group meeting is that others will be able to make use of this technique in their own work...so be sure to ask questions!

Here are the slides.

March

5 Mar.

Karl Young, visiting us from UCSF, will talk to us about "Complexity Measures in Brain Imaging".

12 Mar.

John Mahoney will talk about a quantum type class algorithm.

This algorithm is a way to calculate word probabilities at longer lengths than one can realistically do using other simple techniques.

19 Mar.

Dave Feldman will present "Two-dimensional information theory".

26 Mar.

Take a break!

April

2 Apr.

Khan Mahmud will talk about "Dynamic generation of Schrodinger's Cat State in Bose-Einstein Condensates".

Those interested should look at the paper.

9 Apr.

Sean Whalen will give us an update on his research.

"Kimura proved that a mutant allele in a finite population either fixates or becomes extinct in all individuals in finite time. The stochastic process is called genetic drift. We recast Kimura's neutral evolution model of genetic drift as a special case of a generalized drift stochastic process of structured individuals with memory. This requires a new and more general definition of fixation in stochastic processes based on information theory. We examine the structural drift and fixation properties of several classes of structural drift models."

Here's a couple links explaining genetic drift, the first is more concise than the second but lacks some interesting history:

genetic drift

genetic drift

16 Apr.

John Mahoney will try to explain why we should believe in negative conditional entropies.

23 Apr.

We will attempt to untangle the entanglement from last time...

30 Apr.

Chris Ellison will give a practice talk for his upcoming oral candidacy exam.

May

7 May.

Benny will present an article on learning and its costs in Drosophila.

14 May.

David Dunn will speak.

"Insects, Trees, and Climate---The Bioacoustic Ecology of Deforestation and Entomogenic Climate Change"

Read the article.

21 May.

28 May.

Dave Albers will be speaking.

June

4 Jun.

Paul Smaldino will present : Dynamic Differences in Exploratory Behavior in Vasopressin-Deficient Rats

Abstract:

The Brattleboro (DI) rat has a defective gene for the production of the neuropeptide vasopressin (AVP). Vasopressin's main physiological function is in water retention, but it is also involved in a number of cognitive and social behaviors. AVP has been suggested to play an important role in the onset of autism spectrum disorders (ASD). The Brattleboro rat provides a unique opportunity to study the developmental importance of AVP. Statistical measures of behavior have come under fire for high environmental susceptibility as well as their lack of strong predictive power. We propose to study the spatiotemporal dynamics of behavior. Ideally, this will lead to the development of a computational model that predictively accounts for the motor differences between Brattleboro and wild-type rats. In turn, this may yield a better understanding of the role of AVP in general, and perhaps suggest a more precise diagnostic measure for early onset of ASD.

Look here for a reference article by Jeff Schank.

11 Jun.

18 Jun.

25 Jun.