on Monday, March 27th, 2017 2:07 | by Weitian Sun
5 measurements on self learning from rover:
6 measurements on self learning form sitter:
10 measurements on self learning from wtb:
4 measrements on long term test from rover:
3 measrements on long term test from sitter (fly stops a lot during test, this result may not be accurate):
7 measrements on long term test from wtb
on Friday, February 24th, 2017 1:48 | by Weitian Sun
After adjusting the laser angle, PI got much improved. But, the test 2 phase shows a lower PI than test 1 in most of cases, the reason for that is still unclear. The following is the result of 9 measurements without drifting.
train1 PI: 0.50
test2: 0.06 ???
on Monday, September 26th, 2016 1:29 | by Axel Gorostiza
in order to test the relevance of different sensory modalities in photopreference and wing-clipping effect, I decided to test some sensory mutants. This is the result of the first experiment.
on Monday, February 8th, 2016 11:38 | by Pablo Martinez
After we got the data of the behavior of the fly with strokelitude, we have made the analysis of 40 datapoints. Those datapoints were randomly picked in the middle of our dataset (10 minutes) and we made a second prediction, taking the points just before the end of the flight. For that we made a double correlation with those two predictions for each line and those are the results:
For the two control lines,
WTB x C105;;C232
UAS-TNT-E x WTB
And for the experimental line (UAS-TNT-E x C105;;C232)
As long as the linear regresion decreases, more impredictable is the fly so its behave like non-linear function. The two controls are suposed to be more impredictable than the control line. There is more decrease in those two lines comparing to the experimental line, although is not too remarkable.
We also made an RMSE analysis, Christian explained it in his post. “RMSE measures just the differences of the absolute points whereas correlation coefficient is rather if the direction and degree of variation correlates (covariates)”
Here we have the plots:
WTB x C105;;C232(control)
UAS-TNT-E x WTB(control)
experimental line (UAS-TNT-E x C105;;C232)
As long as the linear regresion increases, the absolute points differ from the prediction and the normal trace.
on Monday, February 1st, 2016 2:16 | by Christian Rohrsen
on Sunday, January 17th, 2016 7:02 | by Christian Rohrsen
These are just 5 flies (WTBxTNT) from the strokelitude where I measured the correlation coefficient on the Y-axis. In the X-axis, first bin is from 0-2 s of prediction, second is 2-4s and so on.
It seems as if some flies do nicer than others. Although it seems to me that a correlation coefficient from 0.3 isnt a big thing with all this variability. I have to find out the best binning though, I think it needs to be much more in the short term.
When I do the mean of the 5 flies measured, I do see a very slight decay. But once more I would say the decay is from the bin 1 to the second.
Here I tried another way, the RMSE, which according to literature and to my own reasoning should be a better analysis. I think RMSE measures just the differences of the absolute points whereas correlation coefficient is rather if the direction and degree of variation correlates (covariates). I find a very weird result. The fit is bad, the it gets better (but it should be just a chance event because correl coef decreases) and then it get very bad and so on.
I think for the future I have to make ensembles of two k neighbours maybe, which seem to increase the prediction power 10-15%. And maybe not look that much into the future as it was done here (10s).
Here some examples of predictions vs observations:
on Monday, December 7th, 2015 2:25 | by Axel Gorostiza
I am trying to spot source of the current behavioral problem we are having. I tested some crosses that I know how they behave. I did the crosses in both directions and tested to different lights sources (cold and warm) at 25°C.
on Monday, August 3rd, 2015 2:58 | by Axel Gorostiza
I started testing the different CS substrains that we have here (CS from the paper http://f1000research.com/articles/3-176/v1 + CS Regensburg). I’ve decided to start with Benzer Paradigm. This is what I have so far.
on Monday, March 25th, 2013 11:51 | by Julien Colomb
It would be great to have some physiolgy data (electrophy) to have a guess about the effect of the PKC in the motorneurons (?)… Do you know if there is a paradigm to look at LTP and LTD at the MNJ?
on Friday, January 18th, 2013 7:04 | by Julien Colomb
I was on a meetup of corporate semantic web last Tuesday. These people are using semantic web technologies (making machine readable content based on ontological terms and relation between these terms) to improve the efficacy of private companies. For instance, they work on ways to improve wiki contents which may be produced in a company. This corresponds at using ontological term to annotate the wiki content and other related technologies. This can be used to find an expert in one category (=somebody who’s posts are rarely corrected on a specific subject).
What is the scientific community (the one which should be leading the way actually) doing during that time: we use text search in “keywords” and titles to find the appropriate literature, that we have to read thouroughly to drive our one conclusions about these different parts… At least, that is what we do 90% of the time, and we all know how inaccurate this can be. Experimental results may be translated into a machine readable content, why aren’t we doing it (it could make everything that much simpler, faster and more accurate)?
The answer: 1. there is no tool nor database where we could do it. 2. Scientists do not have the time to do it, they are over-pressurized to produce data, not to make it reusable or machine readable.
How to push people to use the semantic web technologies, how to ease this use, should it be done by the authors or by the community, pre or post publication, what ontology tool to use,… What can we do? Is anyone asking these questions around? Does a platform like researchgate be a way to introduce this, or should we go for a public solution, inside pubmed for example?
Is any of you asking/answering these questions?
By the way, this post is tagged by none-ontological terms, a shame?