Back from China

Went to China at the start of the Summer to visit with Shulan’s family in Nanjing, and also took a trip to Beijing for a bit of sight seeing. This was just after the major earthquake in Chungdu, so as you see it has been some time since I got back, but wanted to post a few notes about the trip.

First some pictures. A lot of these are from our trip up to Beijing to do some sightseeing. We took an overnight sleeper train up north to get there, which was fun, haven’t done that since I was a child. But I’m sure the enjoyment factor would fade quickly after the novelty factor wears off. We got to see some of the Olympic buildings constructed for this summer, including the new main stadium known locally as the Bird’s Nest. We went and saw many of the historical sights while in Beijing, including the Forbidden City, the Great Wall and the Ming Tombs, and of course since I was with Shulan sampled many of the authentic cuisines of Beijing, searching out the locals favorite restaurants and foods.

As you can imagine I was following the events of the earthquake quite closely. There was an amazing outpouring from the Chinese people while we were there. And the openness and effectiveness of the organized government response was equally impressive. Hopefully their openness signals some permanent future trends, and will not revert to old ways say after the Olympics. I know for example the criticisms by especially the parents of possible irregularities in construction techniques, especially for school and other public buildings, are beginning to be discouraged and quieted more and more by the Chinese officials. We will see if they can close the barn door after letting the horses of progress out this time.

Some of my favorite coverage of the Chengdu Earthquake came from NPR. I am a listener and supporter of public radio, and they happened to have some of their radio reporters in Chengdu for a feature on China focusing on Schizuan providence when the earthquake struck.

Spent the day yesterday at the “Happiest Place on Earth”(tm). A few of our better pictures.

Here I am with good buds Walt and Mickey, posing in front of the Magic Kingdom.

Shulan and myself shackled together in FrontierLand (this is how I imagine we’re destined to end up sometime in the future).

Here’s Shulan looking pretty cute in the Robinson Crusoe treehouse.

Day 5

Wolf Singer gave a wonderful plenary talk this morning entitled “Oscillations, Synchrony and Temporal Codes in Cortical Processing”. Among many good experiments and ideas, I’ll just pick out one.

One of Singer’s main points is that the current debate about the importance of rate coding vs. temporal coding hypothesis in brain neurodynamics is misplaced. Singer argues that both are in fact used, and important. Can’t help but point out a connection with the current book I am reading, Scott Kelso’s The Complementary Nature. In general, Kelso says that many of these controversies and viewpoints, such as Nature~Nurture, Wave~Particle, Time~Space, are really complementary pairs. Viewpoints that stress one as more important than the other almost always turn out to be fundamentally wrong. Deep scientific advancements are often in fact the results of a synthesis (or even a Kuhnian type paradigm shift) that incorporates both seemingly dichotomous forces, and reconciles the seeming incompatibilities.

So back to Singer’s points about Rate Coding ~ Temporal Coding views, I wonder if there are any neurodynamical models that seek to reconcile and represent both in the same model. Most work I am familiar with use either a model like a spiking neuron model, that will necessarily emphasize temporal coding, or maybe a population model, like the K-sets, that focus in on rate representations.

Here are a couple of pictures of myself from the opening conference welcome:

Shulan arrived here in Orlando today. We’re planning on spending Saturday at Disney World, for a brief fun weekend getaway. I’m very excited and really looking forward to going! Its quite a contrast, the type of intense concentration of a conference for the week, then going to an amusement park. Kinda the opposite ends of the spectrum of experience. I guess that is why conferences are often held in such places (or at least the excuses we tell ourselves on why such conferences should be held in such places :-). Anyway I’ll definitely have to post some pictures from our Disney adventure this Saturday.

Day 1

Arrived the other day for the 2007 International Joint Conference on Neural Networks IJCNN Conference. Boy this is an expensive conference! And I really don’t understand these so called “Tutorials” for which the privilege of attending you are supposed to pay an extra $130. Really, these are just another way for these people who proposed and give the tutorials for presenting and communicating their work (almost always just the same thing as an extended special session). They are almost never really tutorials, for the purpose of the participants learning a new skill or tool. One hint to other conference goers, these kinds of tutorials, in my experience, are not worth paying the money to go to them (and in all conferences I have attended, no one actually checks your registration, so you can walk in and listen to them anyway).

Well at least IJCNN provides free wireless for the conference participants, unlike some other conferences I can think of (CogSci I’m looking at you).

Day 4

The Immediate Interactive Behavior symposium this afternoon was quite interesting. It is something of an amalgam, of 3 separate movements in cognition. Embodied and embedded cognition, dynamical systems and dynamics of cognition, and the Cognition in the Wild notions of Edwin Hutchins where we analyze cognition as an interactive dynamical system between environment and person, and especially how we offload our cognitive processes and reduce the cognitive load into our environments by structuring our environments in such a way to facilitate this. This is stuff I am very familiar with, and am quite happy to see it all being combined together into one view. Really this kind of combined view is becoming much more popular. I can name a half-dozen books, like Andy Clark’s Being There, or Rolf Pfeifer’s Understanding Intelligence of which this combination is central to the thesis of how real biological cognitive performance functions.

For the most part loved the seminar on Complex Systems concepts and their application/use in the Cognitive Sciences. This is basically the conceptual perspective I have had, at least since the beginning of my graduate student days. And I was quite surprised, and glad, to actually see an explicit seminar on these topics and their application to Cognitive Science. (This was a first for me at a CogSci confernce, and since I have been going for the last 5-6 years, probably the first time ever for the conference). Unfortunately there were some less interesting subjects (to me), especially some of the presenters were interested in the teaching and learning of complex systems concepts, rather than their application to models of cognition. But besides that it was a very good session.

We had dinner with Xiaowei Zhao and his wife Shuxia the night before. Here is a picture of all of us from their camera. Zhao is a Postdoc with Ping Li at the University of Richmond and their paper won the Best Computational Modeling prize in the language category this year for the conference.

Day 3

Attended John Laird’s plenary talk this morning titled “Is Cognitive Science the Right Method for AI”. John Laird is of course best known for his own cognitive architecture SOAR. I wonder, since John Anderson was invited, they had to, or someone insisted that John Laird also be invited for a plenary talk (or vise versa). Anyway, one quick thought. In general I am encouraged by so called Symbolic Cognitive Architectures in one sense, they are examples of integrative cognitive science, or studying complete cognitive agents and performance, not isolated cognitive functions. In general I believe this is a good thing to be doing for Cognitive Science. Of course, I have issues with their basic assumptions and approach, e.g. starting from a disembodied, symbolic, top-down framework and assumptions.

Bob Glushko made a remark in his Semantics in the Wild seminar that this kind of research (presented in this symposium, tagging & social network systems) hasn’t been seen in cognitive science before. If true the Cognitve Science Society as a whole is a bit behind the curve. Though I might disagree a bit. I do know of at least some very good research by Psychologists / Cognitive Scientists into social networking systems in general. Especially phenomena like Facebook and Myspace, and relations to teen development, culture, etc. Kind of a new type of sociology / anthropology for the technology age. (Though I’m blanking on the 1 or 2 top names I’ve heard of doing this stuff.) But admittedly this symposium’s topic is a bit different. It is looking at semantic tagging in such social systems, and how they naturally evolve (and self-organize), and relation to natural human categorization and concept formation. Some interesting stuff. Especially if you, like me, find this whole idea of leveraging crowd intelligence from social intelligence, a type of borganism concept (a term coined by Charlie Stross I believe).

I thought that Jeff Elman gave an excellent talk in his 2007 Rumelhart Prize Plenary address. His receiving of the prize is richly deserved.

Day 2

A fairly good day at the conference today, though I must admit that I still think the conference is a bit too dominated by people with Psychology backgrounds, and not enough representations from the other disciplines. Cognitive Science has always been hailed for its highly interdisciplinary nature, but there is a dearth of other disciplines lately (AI, neuroscience, etc) at the conference. Though there are new prizes this year for the best computational models of various cognitive processes or areas, a good development. But I find things exemplified by people like the morning Plenary speaker Walter Kinstch (Latent Semantic Analysis) or the evening Plenary speaker John Anderson and his ACT-R cognitive models (a symbolic / expert system type of approach, but with low level mechanisms to model and account for timing and other empirical brain data).

Anyway, there was one particular find (for me) today, a paper by Anthony Morse and Tom Zimke from the University of Skovde on cognitive robotics and enactive perception. A bit tough to get the idea across in a few sentences, but they are using so called reservoir systems (such as Liquid State Machines or Echo Nets) as a way to set up complex dynamic that can be perturbed by incoming (an internal) stimuli. The basic idea is that the combination of the complex dynamics of the system, perturbed by the inputs actually simplifies relations in the sensory/motor associations, and makes them easier to identify. They use the analogy of dropping a stone into a pond of water, the ripples afterwords represent the complex dynamics of the perturbing event. This is really quite similar to Freeman’s (and me and Kozma’s) viewpoint of how the K-models dynamics function in the olfactory system to enable active perception and categorization of smells (and associated behavior). Freeman and I have often used the metaphor of snowflake formation, in the right conditions the atmosphere is in a highly unstable state far from equilibrium. The introduction of an perturbance (a dust particle for example) causes a reaction to catalyze e.g. the snowflake crystalizes. This crystalization is a kind of complex dynamics that encodes the original perturbance in a new way.

The K-models share similar properties with such reservoir systems. They are sparse randomly interconnected networks of artificial neurons. The networks create a natural, very high-dimensional dynamical state-space. The high-dimensional state space can have different inputs (and input sequences) cast into them, and the dynamics react in different ways to the inputs. Unfortunately (for me), I think Morse is way ahead of demonstrating this type of dynamics in a complete cognitive agent, and how it might work to yield active, situated, embodied perception.

Well we successfully arrived last night at the Gaylord Opryland convention center & hotel for the 29th annual Cognitive Science conference. Actually the Gaylord center would be right at home in Texas, it is HUGE (apparently there is a Gaylord in Grapevine, but I haven’t been there, wonder if it is the same thing). It seems like the hotel can’t decide whether it wants to be a theme park or maybe just a mall. At the risk of sounding a bit snobbish, the overall effect is a bit tacky, but the facilities for the conference appear to be more than adequate.

A few pictures from our first day…

Day 1

Went to David Noelle’s tutorial on Computational Cognitive Neuroscience this morning. Mainly a tutorial on PDP++/Leabra, a neural simulation software suite. I was not familiar with the software as I had only heard of it and not used it before. I was impressed. It appears to support many of the same properties of the K-models; including attractor dynamics, like periodic, point and maybe even chaotic dynamics, and using layers for excitatory-excitatory, excitatory-inhibitory, inhibitory-excitatory, inhibitory-inhibitory connectivity, similar to KII and KIII setup and dynamics. It even appears they think of units in terms of representing a homologous (small) population in order to address rate coding model problems, again similar with if not the same as K models. Also use mainly Hebbian correlational learning, as well as an error correction which is probably similar to the weight normalizations done in many K models. However, the emphasis is still definitely on cognitive models that employ point attractor dynamics to form representations, an important difference from Freeman’s K-sets.

Well looks like I’m going to be going to 2 conferences in August here, though I don’t have a paper or talk at either one. From Aug 1-4 I’ll be in beautiful downtown Nashville at the 29th annual Cognitive Science Conference. Then from the 12th to the 17th its off to beautiful Orlando FL for the 2007 International Joint Conference on Neural Networks (IJCNN). Both look exciting, though especially IJCNN (for me) as a lot of our Kozma/Freeman cognitive neurodynamics group are giving talks and tutorials there. If you are going to be to either one, look me up. I’ll probably post a few blogs about impressions, etc. from the conferences while I’m there.