Tuesday 29 November 2011

More privacy stuff

Been lecturing about privacy this week - very interesting as it not only forces you to clearly describe what you already implicitly know but also provides a way for you to expand on your knowledge. I think it was Feynman who said that if you can't explain it simply then you don't understand the subject.

I'm still struggling with this, but we now have a case study which fits within this framework...just need to explain in simple terms the diagram below first:

which should give the intial bits of framework in my "theory of privacy"....the above misses out a few things which I haven't been able to formally explain yet. However just getting this far (even if I don't explain the meanings of O and P here) is quite a leap in conceptualisation of the subject. Clearing out and scoping the concepts leads to a better understanding and now I just need to explain it a few times and I think it might just work....

Saturday 26 November 2011

Database Privacy

Interesting stuff:

Database Privacy: Overview (Microsoft Research)

The problem of statistical disclosure control—revealing accurate statistics about a population while preserving the privacy of individuals—has a venerable history. An extensive literature spans multiple disciplines: statistics, theoretical computer science, security, and databases.  Nevertheless, despite this extensive literature, «privacy breaches» are common, both in the literature and in practice, even when security and data integrity are not compromised.


Friday 25 November 2011

Visual Modelling and Ontologies

Just some papers with the Visual Modelling Group in Brighton:

Howse, J., Stapleton, G. and Oliver, I. (2010) Visual reasoning about ontologies In: International Semantic Web Conference, November 2010, China.
We explore a diagrammatic logic suitable for specifying ontologies using a case study. Diagrammatic reasoning is used to establish consequences of the ontology.

Oliver, O, Howse, J., Stapleton, G., Nuutlia, E and Torma, S (2009) Visualizing and Specifying Ontologies using Diagrammatic Logics In: Proceedings of the Australasian Ontology Workshop, 1 December 2009, Melbourne, Australia.(Best Paper)
This paper proposes a diagrammatic logic that is suitable for specifying ontologies. We take a case study style approach to presenting the diagrammatic logic, and draw contrast with RDF graphs and description logics. We provide specications of two ontologies and show how to depict instances. Diagrammatic reasoning is used to show that an instance conforms to a specification. We also include examples to show how diagrammatic rules can be used to (a) constrain ontology specifications and (b) define mappings between ontologies. The framework also allows the specification of queries. The positive features of the diagrammatic logic are discussed, supporting a claim that the new logic is intuitive and appropriate for ontology specification. Finally, we discuss the possibilities for developing tools to support the use of the diagrammatic logic, including automated diagram drawing and reasoning procedures.

Howse, J., Schuman, S., Stapleton, G. and Oliver, I. (2009) Diagrammatic Formal Specification of a Configuration Control Platform In: 2009 Refinement Workshop, 3rd November 2009, Eindhoven.
This paper presents a diagrammatic logic framework that is suitable for use in formal specification and for reasoning about and refining formal software models. We take a case study style approach to presenting the framework by developing, in some detail, an abstract model for a transparent configuration control platform. The model is built up by stages, corresponding to separate concerns of configuration control. Each successive level is a refinement of the previous level. We discuss the possibilities for developing tools to support the use of the diagrammatic logic, including automated diagram drawing and reasoning procedures. Our wider goal is to make a formal specification easier for its clients to understand.

S Törmä and E Nuutila are from Aalto University, Finland.

Keywords: Modelling, Modeling, Visual, Diagram, Diagrammatic, Logic, Semantic Web, Ontologies, Reasoning, Ontology, Formal Specification

Erdos Number

Reference to an earlier posting on my Erdos Number being 5, well a co-author of mine* - Ora Lassila - has an Erdos number of 3, therefore mine being at most 4...

*this is the one I am sure about, I think my colleagues from the University of Brighton Visual Modelling group also provide a similar length path to Paul Erdos....

The Axiom of Choice and Mathematical Proof...(humour!)

Probably the most controversial axiom in mathematics, but quite simple really...as XKCD explains :-)



Proof by Intimidation needs to be added to the canonical list of "Proof by..."
Aside: an an undergrad many years ago I read a fascinating paper on reactive systems and after 3 or 4 pages of dense, complicated mathematics there was the statement: Proof left as exercise to reader.....it gives you some idea of the calibre of the author that he could get away with that (and no, I'm not telling who....)
anyway, a quick search using an popular search engine reveals a list which I'll just put a few choice examples:

Proof Techniques:

Proof by example
The author gives only the case n = 2 and suggests that it contains most of the ideas of the general proof.
Proof by intimidation
``Trivial'' or ``obvious.''
Proof by omission
``The reader may easily supply the details'', ``The other 253 cases are analogous''
Proof by obfuscation
A long plotless sequence of true and/or meaningless syntactically related statements.
Proof by personal communication
``Eight-dimensional colored cycle stripping is NP-complete [Karp, personal communication].''
Proof by reference to talk
``At the special NSA workshop on computer vision, Binford proved that SHGC's could be recognized in polynomial time.''
Proof by reference to inaccessible literature
The author cites a simple corollary of a theorem to be found in a privately circulated memoir of the Icelandic Philological Society, 1883. This works even better if the paper has never been translated from the original Icelandic.
Proof by flashy graphics
A moving sequence of shaded, 3D color models will convince anyone that your object recognition algorithm works. An SGI workstation is helpful here.
Proof by misleading or uninterpretable graphs
Almost any curve can be made to look like the desired result by suitable transformation of the variables and manipulation of the axis scales. Common in experimental work.
Proof by vigorous handwaving
Works well in a classroom, seminar, or workshop setting.
Proof by cumbersome notation
Best done with access to at least four alphabets, special symbols, and the newest release of LaTeX.
Proof by abstract nonsense
A version of proof by intimidation. The author uses terms or theorems from advanced mathematics which look impressive but are only tangentially related to the problem at hand. A few integrals here, a few exact sequences there, and who will know if you really had a proof?
Disproof by ``not invented here''
We have years of experience with this equipment at MIT and we have never observed that effect.
Proof by personal communication I remember as being something like:
"I met Knuth/Scott/Gödel/... in the corridor the other day and he thought it sounded ok"...

Note the the "Proof by Intimidation" above is slightly different to the XKCD example....both work....hmmm, might try this in my talk on the Theory of Privacy next week....did I tell you about that? Later....

I guess we could also add the "Proof by UML" from Software Engineering, this is explained in the excellent Death by UML [1]  paper.

Proof by Abstract Nonsense.....did someone mention category theory? ;-)


[1] Alex E. Bell. 2004. Death by UML Fever. Queue 2, 1 (March 2004), 72-80. DOI=10.1145/984458.984495 http://doi.acm.org/10.1145/984458.984495

Wednesday 23 November 2011

BBC Research and Development Labs

Very interesting article on The Register about the BBC's R&D Labs and a very good point about industrial-academic research and its value, I quote this paragraph:


Some may question why the BBC puts so much effort into R&D. But with around 50 active projects, and an income of around £16m, including money from the licence fee, commercial licensing – the BBC has over 160 patents – and grants, the R&D team punches well above its weight.


A very salient point: it costs £16m but generates not just patents and their associated licensing income but the technology that drives not only your current TV, Radio and digital media experiences (for a start!) but is defining the future of those. Sixteen million pounts is remarkably cheap for what you're getting which over the course of time is worth many, many times this.

Now with the EPSRC Research Impact Assessments comming up in the UK, this point needs to be hammered home regarding impact and value versus cost.

Tuesday 22 November 2011

EU, Water, Dehydration

The headlines (at least in the UK) screamed: "EU SAYS WATER IS NOT HEALTHY." complete with hysterical, ill-informed, rabid anti-EU comments...

The REAL story behind this however is something very, very different and very serious...I'll let a quotation from a report in The Guardian explain:

Daft hysteria over the EU's ruling on water and dehydration 

Martin Robbins, 18 November 2011

The EU has not said that water isn't healthy, and it's ruling on the vexatious claim that bottled water can prevent dehydration is perfectly sensible

There are two major problems with the claim: drinking water doesn't prevent dehydration, and drinking-water doesn't prevent dehydration. 

Firstly, "regular consumption" of water doesn't reduce the risk of dehydration any more than eating a pork pie a day reduces the risk of starvation. If I drink half a pint of bottled water while running through a desert in the blistering sun, I'll still end up dehydrated, and if I drink several bottles today, that won't prevent me from dehydrating tomorrow. The key is to drink enough water when you need it, and you're not going to get that from any bottled water product unless it's mounted on a drip. 

Secondly, dehydration doesn't just mean a lack of water, or 'being thirsty'; electrolytes like sodium are important too. If salt levels fall too far, the body struggles to regulate fluid levels in the first place. That's why hospitals use saline drips to prevent dehydration in patients who can't take fluids orally, and why people with diarhhoea are treated with salt-containing oral rehydration fluids. Presumably the next big investigation at the Express will expose the shocking waste of NHS money on needless quantities of saline solution, when jolly old tap water would work just as well.

and then refer to the EFSA report:

Scientific Opinion on the substantiation of a health claim related to water and reduced risk of development of dehydration and of concomitant decrease of performance pursuant to Article 14 of Regulation (EC) No 1924/2006

EFSA Journal 2011;9(2):1982 [7 pp.]. doi:10.2903/j.efsa.2011.1982

Abstract
Following an application from Prof. Dr. Moritz Hagenmeyer and Prof. Dr. Andreas Hahn, submitted pursuant to Article 14 of Regulation (EC) No 1924/2006 via the Competent Authority of Germany, the Panel on Dietetic Products, Nutrition and Allergies was asked to deliver an opinion on the scientific substantiation of a health claim related to water and reduced risk of development of dehydration and of concomitant decrease of performance. The scope of the application was proposed to fall under a health claim referring to disease risk reduction. The food, water, which is the subject of the health claim, is sufficiently characterised. The claimed effect is “regular consumption of significant amounts of water can reduce the risk of development of dehydration and of concomitant decrease of performance”. The target population is assumed to be the general population. The Regulation (EC) No 1924/2006 defines reduction of disease risk claims as claims which state that the consumption of a food “significantly reduces a risk factor in the development of a human disease”. Thus, for reduction of disease risk claims, the beneficial physiological effect results from the reduction of a risk factor for the development of a human disease. The Panel notes that dehydration was identified as the disease by the applicant. Dehydration is a condition of body water depletion. The Panel notes that the proposed risk factors, “water loss in tissues” or “reduced water content in tissues”, are measures of water depletion and thus are measures of the disease. The Panel considers that the proposed claim does not comply with the requirements for a disease risk reduction claim pursuant to Article 14 of Regulation (EC) No 1924/2006.


So, once again we read crap in newspapers and find hidden underneath a very serious scientific message...

...in a similar vein, Euromyths contains a vast array of these....well worth reading.

And if you don't like what Europe is doing you might like to consider actually talking to your MEP or even voting for one - at the end of the day, the people of the EU vote for these people in a free and fair democratic process; and, as a citizen of the EU you are entitled to stand as a candidate for the European Parliament



Unintended Consequences in Academia

Found this via a posting on Facebook; a very interesting look at the unintended consequences of trying to measure the output of academics, and just like any other metric the issues that come up on how that particular metric gets misused. Sort of reminds me of the classic software engineering metric leading to bonus payments: bug fixing....

Perverse Incentives in Academia

 In particular I quite like these two:

Researchers rewarded for increased number of citations. Researchers do work that is relevant and influential. H-index obsession; list of references no longer included in page limit at many conferences.

Teachers rewarded for increased student test scores. Improve teacher effectiveness. Teaching to the tests; emphasis on short-term learning.

Of course, one issue here is that you are trying to measure some seriously intelligent people and here's a very good discussion on managing intelligent people by Scott Berkun - I'm not sure he has all the answers but just acknowledging this is a step to proper understanding.

Finally, a nod to that classic tome of unintended consequences (and the economic theory that drives this): Freakonomics by Levitt and Dubner.





Saturday 19 November 2011

Higgs by Christmas?

Maybe Christmas 2011...Nature has an article - Higgs Hunt Enters Endgame - about analysing the final set of data that may, or may not, prove the existance of Higgs:

I'd like to have posted a little of the article, but while (and I applaud this!) Nature allows a small text extract, it does require that I sign up to Rightslink ... so much for citation and reasonable use...


New Physics?

First hints of "the new physics"?  At least these results are becoming more public but still not at the statistical level where they become official, however news of CP violation at this magnitude is extremely interesting. First the BBC's report on this (readable by the layman):



LHC reveals hints of 'new physics' in particle decays
15 November 2011 Last updated at 12:18 GMT
By Jason Palmer Science and technology reporter, BBC News

Large Hadron Collider researchers have shown off what may be the facility's first "new physics" outside our current understanding of the Universe. Particles called D-mesons seem to decay slightly differently from their antiparticles, LHCb physicist Matthew Charles told the HCP 2011 meeting on Monday.  The result may help explain why we see so much more matter than antimatter. The team stresses that further analysis will be needed to shore up the result. At the moment, they are claiming a statistical certainty of "3.5 sigma" - suggesting that there is less than a 0.05% chance that the result they see is down to chance. The team has nearly double the amount of data that they have analysed so far, so time will tell whether the result reaches the "five-sigma" level that qualifies it for a formal discovery

and the report from CERN itself:

Charming surprise
The Bulletin - Issue No. 45-46/2011 - Monday 7 November 2011

The CP violation in charm quarks has always been thought to be extremely small. So, looking at particle decays involving matter and antimatter, the LHCb experiment has recently been surprised to observe that things might be different. Theorists are on the case.

The study of the physics of the charm quark was not in the initial plans of the LHCb experiment, whose letter “b” stands for “beauty quark”. However, already one year ago, the Collaboration decided to look into a wider spectrum of processes that involve charm quarks among other things.

The LHCb trigger allows a lot of these processes to be selected, and, among them, one has recently shown interesting features. Other experiments at b-factories have already performed the same measurement but this is the first time that it has been possible to achieve such high precision, thanks to the huge amount of data provided by the very high luminosity of the LHC. “We have observed the decay modes of the D0, a particle made up of a charm quark plus a u antiquark”, explains Pierluigi Campana, LHCb Spokesperson. “In particular, we have studied and combined the decay rates of the D0 and its antiparticle. According to the theory of the Standard Model, we should have measured a very small value of a parameter known as Delta ACP that is calculated using these decay rates and is related to the properties of matter and antimatter. We found that Delta ACP is around 0.8% instead of the predicted 1‰ (or less). Although making precise evaluations in processes involving charm quarks is difficult, the Delta ACP parameter appears to be much higher than expected”.

Finally, Ars Technica has an overall write-up on the FTL neutrinos, CP violation and Higgs:



It has been a busy week in the world of particle physics, with attention focused on the home of the LHC: CERN. This year, the LHC generated five inverse femtobarns worth of data—nearly half the amount generated during the entire lifetime of the Tevatron—before shutting down the proton program a few weeks ago. From now until its scheduled winter shutdown, the LHC will be doing lead ion collisions to examine the quark-gluon interactions that dominated the Universe immediately after the Big Bang.

Friday 18 November 2011

Fishing

Fishing at Sunset

Wierd Tales

The Guardian has two interesting stories today, one fiction's unsolved mysteries and the other on "wierd" stories. This latter one has its own website complete with various tales...wierd.

Neutrinos

Neutrinos are wierd things (see here) ... and the faster-than-light results from last month have been reanalysed and new experiements taken place...with the same results.

Possibly, there are still some questions regarding gravitational effects but this second set of experiements is still showing the same results.

Now, I wonder when Higgs will turn up?





Tuesday 15 November 2011

Voyager 2 and LHC

Been a bit slow this month with things to write - actually plenty to write but little time to actually write here. Anyway, two from that bastion of tech literature, El Reg...

 The first is about the current results from the LHC and also an exercise in how to write a science article for the layman:


LHC results may solve riddle of how universe can exist

Antimatter bombshell set to explode physics applecart

Top boffins at the Large Hadron Collider – mightiest particle-punisher and largest machine of any kind ever assembled by humanity – say that they may have uncovered a vital clue explaining one of the greatest mysteries of physics: namely, how is it that matter itself can exist?

Good to see the Standard Model getting a bit of a run for its money...one paragraph spurred a superb retort in the comments:


One of the things the Hadron Collider can do by means of blasting protons into one another head-on at just a gnat's chuff less than light speed is create all sorts of very rare and exotic particles – and the antimatter versions of themselves. Almost anything you might want in the way of crazy particles will appear in the shattered sub-subatomic wreckage spraying out of the proton pileups on the Collider's 27km underground orbital motorway.
and the retort from "a non e mouse":
"Almost anything you might want in the way of crazy particles will appear..."
Except, maybe, the Higgs Boson

However it does look like we might have a lightweight Higgs....maybe....

and then over to the still running Voyager probes and in this case the efforts by NASA to keep Voyager 2 running for another 10 years...(El Reg writes space articles in a similar innuendo filled style as LHC articles...)

Voyager 2 finally agrees to a long hard thrust

Probe takes light-ages to return boffins' calls

Voyager 2 has finally gotten back to NASA to let engineers know that its switch to back-up thrusters was successful.
The space agency sent the signal last week to advise the old explorer to switch to back-up thrusters in order to conserve energy so it can continue its voyage for another decade.
 
Not bad when you consider Voyager 2 was launched in August 1977 and a touch over 9 billion (9,000,000,000) miles away...
 

Saturday 5 November 2011

More Satans Rormokare

Nokia Rocks happened and here's the result....



First public performances of Disaster and RIP.  Starting at 25:00 is my John Entwhistle impression :-)