P = atan(x) with pi/2 > P > -pi/2
http://www.mathworks.com/help/techdoc/ref/atan.html
but P = atan2(x,y) with pi > P > -pi
http://www.mathworks.com/help/techdoc/ref/atan2.html
which makes all the difference in the real world (which is not limited between +-90 degrees, but can do the full circle)!
A good way to estimate an average angle, A, from a set of angle measurements a[i] 0<=i
sum_i_from_1_to_N sin(a[i])
a = arctangent ---------------------------
sum_i_from_1_to_N cos(a[i])
Note, it’s not atan, but atan2!
See here e.g. <http://en.wikipedia.org/wiki/Circular_mean>
from <http://stackoverflow.com/questions/491738/how-do-you-calculate-the-average-of-a-set-of-angles>
Tracter is a data flow framework. It was written initially for doing speech signal processing in the context of Juicer. It has developed into more of a wrapper for other packages. For instance, it can generate HTK features by linking with HTK’s library. Written and maintained at Idiap, tracter contains contributions from other AMIDA partners at the universities of Sheffield and Brno.
Seven years ago, a group of researchers set out to design a computer programme that could observe meetings and interpret conversations and staff behaviour. This programme would realise when staff were getting bored, take meeting minutes and allow users to search recorded conversations. Join Mike Lincoln from the School of Informatics as he explores the impact of this fascinating project.
Book here <http://www.sciencefestival.co.uk/whats-on/categories/talk/non-bio-boom-the-walls-have-ears>
- EADS produces world’s first bike using revolutionary ALM technology – ‘grown’ from high-strength nylon powder
- Called the ‘Airbike’ because Airbus was the first EADS company to use the technology
- New technology will transform manufacturing around the globe
Knowles is the leader of MEMS Microphone. The commercial field grows rapidly, lead by Akustica and Knowles. With the growing market, new companies start to enter the competition. As market leader, Knowles must do everything to protect the intellectual rights.
The following document is a summary of news articles related to the MEMS Microphone filed, in particular, the patent and technology dispute between major players, Knowles, MEMSTech, Analog Devices (ADI), and Akustica. The key disputes are focused on US Patents 6,781,231 and 7,242,089
On the court, Knowles was able to fend off MEMSTech. In November 2010, Knowles learned that the court agreed that ADI violated its patents, but then got disappointed by the court’s decision that these patents were invalid.
Chang Liu (www.MEMSCentral.com)
Written December 8, 2010
The following information is presented in chronicle order.
read here: <http://memscentral.com/Secondlayer/mems_microphone_patent_case.htm>
AN innovative research and development initiative between hi-tech security venture Cassidian, Cardiff University and the Welsh Assembly Government could be replicated in Scotland.
Addressing a meeting of Cardiff Breakfast Club yesterday, head of capability at Newport-based Cassidian, Tony Bagnall, described Foundation Wales as an exemplar for aligning research and development with the needs of industry.
Cassidian was formerly EADS DS and part of global group EADS.
The foundation, which was launched in 2009, last year generated research with a value of more than £5m.
Mr Bagnall said: “It aims to grow and strengthen regional industrial and academic capabilities, which is of such strategic importance here in Wales.“We have to put our intellectual property, which is independently valued, or hard cash into the foundation. And we will collaborate amongst our own research scientists with academia.
“However, the great thing is that SMEs and other organisations that have a niche technology or idea – who are ordinarily swamped by a big corporate organisation – can use the foundation… which is a very soft place for these folks to lie.”
Newcastle-born Mr Bagnall said any intellectual property generated by the foundation didn’t belong to Cassidian, the Welsh Assembly or academia, but the foundation itself.
He added: “If we then want that IP we have to buy it, with that cost going back into the foundation to generate more research.
“Last year the foundation generated research of around £5.2m, which in the current climate is quite phenomenal.”
He said the foundation was strongly aligned to the needs of industries and was a model which the Scottish Executive was looking to replicate.
“We are delighted that we have pioneered this here in Wales and that the Scottish Executive is now looking at it.”
He said Cassidian was continuing to attract the very best graduates into Wales. He added: “They want to come to Wales and work for us and that is partly because we have an interesting graduate entry scheme. If you are an engineering graduate it takes 18 months to go around six divisions within EADS to get the deep and rich flavour of our broad portfolio of engineering.
“We have a very high retention rate of graduates, which is not necessarily because of the state of the economy at the moment, as this has been going on for the last 20 years I’ve been associated with the company.”
He said the focus was also on engaging with school children.
However, he said the company has been affected by the impact of the Government’s immigration cap in terms of recruiting staff outside of the European Union.
He said the world of academia had made a “loud case” in terms of the negative impact of loss of fee income from a reduction in non-EU overseas students, but that it was also impacting on industry.
He added: “In terms of our internships we can take as many as we like from the EU, but the cap on students outside just isn’t practical.”
He said Cassidian had attracted interns from India and Singapore. He added: “They decided to come to this division in Wales, which is phenomenal and a credit to the country. However, because of this cap, we could not take them on now as interns.
“EADS is a large company and we found space for them in other divisions overseas.
“However, when the first wave came to us and went back to their universities, they spread the word of how good a country Wales is, how good a company [Cassidian] is. However, this is now an issue.”
He said the Government cap hit so quickly there was no time to make alternative arrangement.
He described inspiration as being the “creators” of inventions that are going on every day.
He added: “In the UK we are superb at inspiration but we are not so good at innovation, which is where we take that creation and do something with it in an entrepreneurial style and then pass it over to the engineers and the production people. They have to sweat and perspire and give you something that you can sell.”
Cardiff Breakfast Club is sponsored by Morgan Cole, Marsh UK, Lloyds Banking Group and the Western Mail
20 January 2011
By Paul Jump
David Colquhoun tells Paul Jump that the only way to maximise UK potential is to restrict research to an elite circle
…
In an interview with Times Higher Education, he argued that when more than 40 per cent of the population enters higher education, the traditional honours degree – which aims to prepare school-leavers for independent research within three years – is too specialist, while the advanced teaching required is beyond the abilities of some academics.
…
from <http://www.timeshighereducation.co.uk/story.asp?storycode=414890>
http://www.timeshighereducation.co.uk/story.asp?storycode=414579
16 December 2010
By Paul Jump
Journal articles are an outdated way of sharing scientific research, says open-access advocate. Paul Jump reports
“Once you start looking at how the scholarly communication system works with any degree of outside perspective, it looks utterly insane.”
This is the view of biophysicist Cameron Neylon, a senior scientist at the Science and Technology Facilities Council’s Rutherford Appleton Laboratory at Harwell, Oxfordshire, and author of the blog Science in the Open.
He said the current system of communicating the results of scientific research via journal articles is a 17th-century solution to a 17th-century problem. “Printing was adopted because researchers got tired of sending letters to each other,” he told Times Higher Education.
“Publishing was essentially letter aggregation. When there became too many letters, peer review was introduced. You can argue that the biggest innovation since then has been the removal of ‘Dear Sir’ from the beginning of articles.”
Dr Neylon believes that if scholarly communication were redesigned from scratch for the digital age, it would look radically different. Most significantly, the monopoly of the journal article would be smashed. He conceded that articles would still have their place, but added that they fail to maximise funders’ return on their investment because they almost never contain enough information to allow other researchers to replicate experiments.
Those who wish to do so are obliged to request the relevant information from the authors, but “compliance is around 20 per cent”.
Worse still, a lot of potentially valuable research outputs that do not “fit into a story” currently either get crowbarred into a paper or, more likely, never see the light of day.
Dr Neylon thinks it would be far better for all the artefacts of the research process, such as videos, samples, data and images, to be made freely available in an open-access format – hosted either by journal websites or alternatives such as university repositories, individual researchers’ websites or large commercial providers such as Amazon.
Nor is Dr Neylon worried by the potential for information overload to which this proliferation of information could give rise.
“The idea that we need to protect ourselves from the flow of information is getting the web completely backwards. Rather than filter failure, we have a discovery deficit,” he said.
The trick is to develop specific search algorithms that allow scientists to find the information they are looking for.
End of the peer show
Dr Neylon said that researchers are “obsessed with their legacy” and value the article system for its ability to identify particular people with particular ideas at particular times, ensuring that “the guy who publishes three days later loses”.
But he thinks the technology exists to authenticate both the date and the veracity of online content, ensuring that “notches on bedposts” can still be recorded.
He admitted it may be more difficult to replicate journals’ role, via peer review, of separating the scientific wheat from the chaff.
“If someone tells me a piece of data is reliable, I’ll pay it more attention if they are experts. We have to be able to make similar judgements in some form at web scale. There are examples of it working at some level but nothing in the research space that demonstrably could replace peer review and (replicate) the confidence people have in it,” Dr Neylon said.
On the other hand, he thinks that confidence is misplaced. He agreed that peer review is the “core of science” but noted there is “more peer review of the efficacy of homeopathy than there is of the efficacy of (traditional forms of) peer review”.
He argued that the traditional three opinions are not statistically significant enough to make “binary decisions” about whether to publish a paper. Far better, he thinks, to publish everything and leave it to readers to make judgements about quality, either through commenting or via metrics such as how many times an article is read or cited or how many times data are reused.
Dr Neylon is an academic editor of the online journal PLoS ONE, which is pioneering such an approach, and selects papers only on the basis of their rigour rather than their importance. But he admitted that, in general, article-level metrics and commenting facilities currently remain rather crude.
Getting there from here
Dr Neylon admitted his brave new world would require a “big cultural shift” currently being resisted by “entrenched financial interests” in the academy and in publishing.
But he sensed a “building momentum in certain areas”. The publisher Elsevier, for instance, is involved in a number of projects to make datasets more available.
And he is confident a tipping point will be reached within the next decade – particularly given the pressure from funders for the impact of research to be maximised.
“Public research funding is not a sheltered housing scheme for people with PhDs,” he said. “It is something that is expected to deliver and communications is part of that.”
paul.jump@tsleducation.com.
Currently reading “Voicing Features for Robust Speech Detection” by Trausti Kristjansson, Sabine Deligne, Peder Olsen, all IBM
Abstract
Accurate speech activity detection is a challenging problem in the car environment where high background noise and high amplitude transient sounds are common. We investigate a number of features that are designed for capturing the harmonic structure of speech. We evaluate separately three important characteristics of these features: 1) discriminative power 2) robustness to greatly varying SNR and channel characteristics and 3) performance when used in conjunction with MFCC features. We propose a new features, theWindowed Autocorrelation Lag Energy (WALE) which has desirable properties.
Get it from <http://www.strokerecognition.com/pederao/eurospeech_voicing_features.pdf>
 | Wendy Chen on Overview of MEMS microphone te… |
meetingdiarisation 