I'm thinking

No, I'm not dead. But as a I proceed to get exponentially more awesome (haha yea right!), maintaining pseudo-anonymity is more challenging. Which means I need to blog about more abstract things rather than puking up what happened today with pertinent names cut out. And we all know how much your d-list monktress hates thinking.

So yea, I'll get back on the bull, eventually

figure 1: The monktress has the thrown, but she will rise again!

Namste, Itches

Hermitage enters lab right, happy and refreshed from her vacation*.

Hermitage: How fares this fine and beauteous morning!?

Greek Chorus: Greetings and salutations, Hermitage.

lab equipment whirs and beeps greetings to Hermitage

Greek Chorus: By the way, Hermitage we could not follow the 2-step process to dispose of Fraking Hazardous chemicals so we shoved it all in this closet. Could you take care of that?

Hermitage: ....

Greek Chorus: Oh, and we broke half of your shit while you were gone. Can you get on making more shit?

Hermitage: THIS IS SPARTA!!!!!! **

Hermitage charges with sword in hand. **

figure 1: I hope you ate your fraking Wheaties, cause it's on now

--End Scene--

*That's right, I took a mf-ing vacation. Because I fraking earned it.

**Ok, fine, I actually just stomped away to get more coffee. And did some under the breath muttering. If they'd gotten Ebola I probably wouldn't have even felt sorry for like 0.002s.

Growing the @#!$ up

Muffins*!

I'm sure you all have been crying into your pillows at night wondering where your most wondrous and favoritest d-list monktress had pranced off to. Well, she was entering chrysalis to transition from a n00b first year PhD student to a beautiful 2nd year PhD butterfly.

AKA being saddled with a shitton of responsibilities that I did not know even existed before now.

figure 1: Accurate representation of Hermitage's new and improved state

I won't say this is an unwelcome transition because it's A) evidence I'm growing up and B) shows my PI has faith that I know my ass from my elbow. I also won't say that I'm overly thrilled with my new duties that I'm pretty sure were express-mailed from hell. Because, as is probably often the case for the rest of the Academic universe, I came into this year thinking I would have have a blissful amount of time to do AWESOME PWNZ0R SCIENCE. And find myself at 6pm wiping my eyes and saying 'where the Sam diggity bullocks did my day go?'**

Yes, I can hear all you elder monks chuckling at my naivete. LAUGH IT UP CHUCKLES, AT LEAST I HAVE RITTER'S CHOCOLATE IN MY DESK. *sticks tongue out*

But really I should just shut up and bear it because all this shit is important to my continued growth as a future member of the Academia Cult. I manage relatively important equipment and communal lab spaces, I'm assisting with classes and doing a little mentoring of students (except when I forget that they exist and make them reschedule their meetings with me umpteen times. I totally still remember when this used to piss me off to no end in undergrad. My response now? Suck it), and I am allowed to handle Super Important Elder Monks with no Awesome PI pre-filtering. Look at that? Isn't that fantastic? I'm sure you're all thinking 'Psssh, shutup Hermitage, you should be glad you don't have to spit shine the glassware!' And you know what my response is? In all it's glorious maturity?

BUT I WANT TO GET MORE RESEARCH DONE DAMMIT, NO FAIR, NO FAIR.

And, naturally, blogging got chucked out the window once I started seeing my week fill up with meetings and classes with me frantically getting off an experiment within those piece of shit 1 hr windows that occur all too frequently when all you want is ONE decent 4 hr chunk of time when you're somewhere close to the top of your game (i.e. after 9am and before 7pm). And I think I'm finally settling down and adjusting to this new level of mayhem (hence the bloggingsauces).

But I'm trying to be grateful. I'm on multiple papers that are under review. I'm going to my first conferences soon (w00t!). I managed to make a dinner that didn't come from a can or plastic-wrap 2 days in a row now. I remembered to wear socks today.

Sleep-deprived Weekend Hermitage threw out 3 days of work because she forgot that it took 3 days to do.

Baby Steps.


*It has a long enough absence to justify my use of favorite reader petname.
** Bitch please, of course I work past 6pm. I'm just saying.

Unscented Kalman Filter for Brain-Machine Interfaces-Update

I have been absent from my blog for quite a while now, so unfortunately I missed this until today. Dr. Zheng Li of Duke University, first author of the PLoS ONE article that Toaster and I nefariously reviewed, was kind enough to stop by the blog and answer some of my questions. I repost them here so they reach a wider (stop laughing!) audience, in case others wanted to know more about their model. Thank you Dr. Li! My questions are in bold, Dr. Li's responses are in italics.

"However, I was a little surprised with the comparisons they used. While the 1st Order Kalman and Wiener filters may be 'standard', I didn't find it in any way surprising that a quadratic model with memory states would outperform them for such a complex task. I would have liked to see a comparison between the 10th order unscented Kalman filter and another quadratic model, or at something similar."

You are correct that the improvement was not surprising. I suppose the surprising part is that such models not been used earlier for BMI prediction. There has been mention of similar quadratic models in the literature, but only for modeling data, not for prediction. We built upon that work and put it together with an efficient filter capable of non-linear operation for this paper.

I have compared, though anecdotally and unpublished, the quadratic model presented in this paper (with the square root term) with a quadratic model that comprises of each feature, its square, and all pairs of products (I'll call it the vanilla model). The vanilla model was slightly more accurate, though the difference may or may not be significant. We chose to publish with the model with square root terms because it has a firm basis in prior work in the field. I also looked at 3rd order models (similar to the 2nd order vanilla model in form). Again, anecdotally, they offer another slight improvement, but the advantage seems to disappear when combined with using history in the model. In other words, the higher order terms may be capturing some of the same information as the history. 4th order and higher models did not seem to improve at all. Of course, take this all as non-rigorous commentary.

"I didn't see a clear description of computational cost for this method, and how that would increase with complexity of the task."

At the end of the results section is a brief report of the actual execution speeds of our MATLAB and C++ implementation. The discussion of the computational complexity of the filters did not make the final edit, due to the journal's probable audience. The two most time consuming operations in the UKF is the computation of the sigma points and the inversion of the predicted observation covariance matrix, both of which are performed using Cholesky decomposition (in our implementation). The standard Cholesky algorithms are O(n^3), where n is the size of the state space (history taps times feature dimensions) in the former and the number of observations (neurons) in the latter. In our data, the latter was typically larger, up to 240. Though, for a realistic robotic arm with something like 20 degrees of freedom and, say, 10 history taps, the n for the state space is comparable.

With regards to computational increase due to the complexity of the task, my guess is that (within the UKF framework presented in this paper) a more complex task would require more history to model, and so a filter with a larger state would perform better, increasing the cost of the sigma point calculation. However, if we believe the neural tuning model's maximum time offset is independent of the behavioral task, a modification can be made to divorce the state space needed for the transition (movement) model with the state space used by the observation (neural tuning) model, avoiding this increase.

"A BMI capable to 'catching' a target is obviously valuable for someone who is paraplegic, this can open an entire world to them via computer screen. However, one would hope to use BMIs for intricate physical tasks, such as operating a robotic prosthesis. I imagine the amount of training required for the model 'history' in order to have passable execution of even mundane daily (physical) tasks would be enormous. I also wondered how well such a model can handle a sharp change in a 'periodic' behavior, which is not unusual in a real-life context."

A mult-joint prosthesis would require many more dimensions in the state space. The modeling of the movement would also need to be far more sophisticated. This paper describes how to deal with one stereotypical task. In the future, I hope to solve the problem of learning and switching between options in a library of such tasks, giving the prosthesis user a repertoire of practiced actions, not unlike the "muscle memory" people have with their biological limbs. This switching between stereotypical tasks is one example of a sharp change in a periodic behavior. More generally, the BMI can fall back on a "generic" model of movement if it can’t understand what the user is trying to do.

I'm also interested in, and working on, ways to minimize training on the part of the prosthesis user, by moving that training burden to the algorithms. Hopefully we can avoid lengthy training to obtain passable execution. Another approach is to off-load some of the control of the prosthesis to an artificial intelligence and have the user issue high-level commands, like “pick up the cup”.

Famous n00b

Who won blog pick of the month on the PLoS ONE community blog?

I DID!

Ok, fine, Toaster did too.

figure 1: This is one of The Death Star's results for 'victory'. The internets are a scary place

Apparently our Nefarious Nerd Wars have caused such utter carnage that the rest of the blogosphere took notice. Thank you Coturnix for hiding in the bunkers to observe our Nerdcular showdown (Nerdcular, Nuclear? Get it, nyuck nyuck nycuk)! We also won Tshirts, which is pretty snazztastic. Except Toaster's getting mine as well because, like a superhero, I remain nameless to all.

Now, if you excuse me, your favorite d-list monktress needs to go e-gloat some more. Possibly with bon bons.

Lab N00blings

It has occurred! The new lab sibling has arrived and unfortunately he is smart, friendly, and dare I say...cool. I guess I can't vanquish him now.

figure 1: The scales of justice have tipped in his favor. Dangnabit!

This decision may have been slightly affected by the fact he now does all the scut work I used to do. This whole 'senior grad student' business has it's advantages. *cue evil giggle*

I'm still shanking him though if Professor Positive starts talking about my Favorite Smashmouth Sport with him.

Victory is MINE, n00b

I have returned from the bloody battlefield victorious against mein Uber Archnemsis Toaster VICTORIOUS.

figure 1: Hermitage totally cosplayed as Saya to celebrate this victory, and in honor of COMIC-CON zomg zomg!

The final score breakdown (for those who are curious)

Hermitage: 205
Toaster: 195

As you can see, it was a bitter battle until the end. Thanks so much to everyone who participated, it meant a lot to us! Give yourselves a gold star for braving our collective mad scientist wrath to express your (eloquent and insightful) opinions.

I discovered several things in reading the comments on our respective posts. The most important were: 1) People firmly believe in my ability to ruthlessly kick their asses (why thank you!) and 2) People found my analysis to be fairly technical.

Which I thought was hilarious because this work is eleventy billion miles from my area of expertise and I thought I wasn't being technical enough. I immediately began to wonder if this is some mind-warping effect of graduate school, that you start thinking that having bigger, fancier words and more tortuous terminology shows you actually understand the paper. When you know, stating it in plain language is actually a greater demonstration of understanding.

I will, of course, be bunkering down for the inevitable Nordic retaliation from Toaster. This is only the first battle in the eternal war for Nerd Supremacy. But for now I will rest (lightly) on my laurels and claim the title of MOST NEFARIOUS. Thank you for the fun times, Toaster, one can only hope to have an archenemy as evil and fluffy as you are^^.