Open Source Computing: Open CL

Here is a great presentation from the Khronos group on the new Apple sponsored open source language.
http://www.khronos.org/opencl/

OpenCL – The open standard for parallel programming of heterogeneous systems”

OpenCL™ is the first open, royalty-free standard for cross-platform, parallel programming of modern processors found in personal computers, servers and handheld/embedded devices. OpenCL (Open Computing Language) greatly improves speed and responsiveness for a wide spectrum of applications in numerous market categories from gaming and entertainment to scientific and medical software. OpenCL supports a wide range of applications, from embedded and consumer software to HPC solutions, through a low-level, high-performance, portable abstraction. By creating an efficient, close-to-the-metal programming interface, OpenCL will form the foundation layer of a parallel computing ecosystem of platform-independent tools, middleware and applications.

OpenCL is being created by the Khronos Group with the participation of many industry-leading companies and institutions including 3DLABS, Activision Blizzard, AMD, Apple, ARM, Broadcom, Codeplay, Electronic Arts, Ericsson, Freescale, Fujitsu, GE, Graphic Remedy, HI, IBM, Intel, Imagination Technologies, Los Alamos National Laboratory, Motorola, Movidia, Nokia, NVIDIA, Petapath, QNX, Qualcomm, RapidMind, Samsung, Seaweed, S3, ST Microelectronics, Takumi, Texas Instruments, Toshiba and Vivante.

Biz Stone finally talks business

Twitter co-founder, and creator of Blogger Mr Biz Stone finally set out a short brief email ( or twemail) on the changes in Twitter’s terms of service.The very concise email is below and an excellent interview with the man is at http://therumpus.net/2009/04/the-rumpus-interview-with-twitter-co-founder-biz-stone/

Incidentally Biz Stone’s inventions are kind of revolutionary in social media – he also founded Blogger ( blogging and micro blogging have done more to confound LarryRank algorithm at Google Search than anyone else).

What does an analytic, data whining blog have to do with social media. Plenty. If you have ever designed a propensity scoring model for targeting customers based on their behavior , more clean data that is identifiable an individual level is always a boon. The current trend for sentiment analysis is simply addition of text keywords ( or categorical variables if you insist) to the existing customer database.

Can adding keywords from blogs, tweets, web searches, TO existing data about you (credit bureau, demographic, purchase behavior)- can this lead to a better lift in the models. Yes.

Will this lead to more privacy debates. Yes. Given the huge volume of text variables, as well as the huge number of potential customers- privacy debates are quite statistically irrational ( but we digress into economics here).

No one is interested in selling just 1 more product. They use people (nicknamed Numerati) for writing queries to append, manipulate data so as to AGGREGATE and then build a model. Only after the models are built are the scores disaggregated AND scored individually- usually in automated manner.

No company is interested in selling to one consumer so they dont stoop at a privacy invasive search of individuals.

Advertsing is not an evil way of making money, Mr Stone. Just Trust Google and the guys who could not complete their Phd because they WERE making money.

What if all maths grads did that- ..and that’s an interesting thought.

Hi,

We’d like to let you know about our new Terms of Service. As Twitter
has evolved, we’ve gained a better understanding of how folks use the
service. As a result, we’ve updated the Terms and we’re notifying
account holders.

We’ve posted a brief overview on our company blog and you can read the
Terms of Service online. If you haven’t been by in a while, we invite
you to visit Twitter to see what else is new.

Overview: http://blog.twitter.com
Terms: http://www.twitter.com/tos
Twitter: http://www.twitter.com

These updates complement the spirit of Twitter. If the nature of our
service changes, we’ll revisit the Terms as necessary. Comments are
welcome, please find the “feedback” link on the Terms of Service page.

Thanks,
Biz Stone, Co-founder
Twitter, Inc.


Presenting R

Here is a short presentation I made for fellow students at work.

It is generally at a beginner’s level or for people having trouble transitioning to R.

and if you want to see the video presentation you can see it from here on UTK’s lecture capture mechanism

Title: R Help Session
Speaker: A. Ohri
Desription: Session for R Beginners
URL: http://vcweb.bus.utk.edu/20090911-103113-cap403/

In addition, here is a link for the handout:

https://docs.google.com/present/edit?id=0AdYMMvghK2ytZGN2c3MzNThfODA2ZzY4N2I5bno&hl=en&invite=CKvforMH

Oracle announces Oracle Database 11g Release 2

Here is the product release site for 11g http://www.oracledatabase11g.com/ As per Oracle communication it continues to be the undisputed leader in big data batabase category.

Some other product features/ claims-

The Oracle Database Machine and Oracle Exadata Storage Server use intelligent storage software to process queries in the storage layer and get answers at least 10 times faster than before.

Oracle Database 11g Release 2 drives down hardware costs by enabling consolidation in a shared computing environment—and reaping the price and performance advantages of commodity hardware. The edition based redefinition feature lets you upgrade applications while users remain connected to their business applications.

oracle

And interestingly

It includes numerous automations and advisors that help double database administrators’ productivity over previous releases.

Claims of 10X, and 100% productivity always enhance the statistical thinking in my head. Did the new software really do so much improvement or was the earlier release just plain daffy. Good software can rarely go up 10X faster and 100% more productivity in a single release.

However, Oracle continues to be number 1 as of now-

Two analyst firms recently published their market share numbers for 2008, both finding Oracle continues to lead the industry, topping Gartner’s 2008 Worldwide RDBMS Market Share Report with 48.9 percent share, and leading IDC’s 2008 Worldwide Data Warehouse Platform Software 2008 Vendor Shares report with 40 percent share.

Interview Stephen Baker Author The Numerati

Here is an interview with Stephen Baker, the author of the famous and remarkable book The Numerati. Stephen is the senior editor at Businessweek and his remarkable book made the world sit up and pay attention because for the first time, anyone wrote of the increasingly quant driven lives we lead thanks to the internet and the analytical brains that power the stimulus, design and targeting of it. Increasing amounts of data is collected about consumers than at any previous point of time in human history and the number crunchers or the quant jocks are the ones who increasingly help with decision making and decision management. Steve calls these people “The Numerati” or the new math people who help shape our lives.

There will always be lawyers and financiers who make loads of money. But they will have quantitative experts on their teams- Stephen L Baker

Ajay- Describe your career journey from high school to a technology writer to author of The Numerati.

Steve- I was always interested in history and in literature, and in college I fell in love with Spanish. So after college, I moved to Ecuador, taught English, and wrote fiction. I saw early on that I wasn’t going to be able to make a living with fiction. So I went into journalism. My goal was to become a correspondent in Latin America. Through my 20s, I worked in Vermont, Madrid, Argentina, Venezuela, Washington DC, and El Paso, Texas. And I finally got the job I was looking for, bureau chief for the Mexico bureau of BusinessWeek magazine.

After Mexico, my family and I moved to Pittsburgh. It appeared that the magazine was losing interest in heavy industry in the mid-90s, so I began to write about software and robotics coming out of at Carnegie Mellon University. That was my transition into technology. A year later, BusinessWeek offered me a job covering technology in Europe. I moved with my family to Paris, where we lived for four years. I focused largely on mobile communications. It seemed to me that the combination of mobility and the Internet would fundamentally change communications.

I returned to New York in 2002. I focused on big picture stories. One day in 2005, I proposed a story about the decline of the U.S. technology industry. I argued that we were behind in wireless and in broadband, we were graduating fewer scientists and engineers than other great powers, especially in Asia. One editor pointed out that mathematics was critical for these competitive issues. The editor in chief, Steve Adler, called for a cover story on math, and he assigned it to me. I didn’t know much about math at the time, and I still don’t. But this gave me the chance to dive into the world of data analysis. I wrote a cover story, Math will Rock Your Business, and later got the contract to write the Numerati.
Ajay- How do you think the government can motivate more American students to science careers?

Steve- I think focusing on the science that kids find cool–robotics, space and ocean exploration, would help. Funding basic research would be useful. But I don’t think it’s entirely a governmental issue. Parents, companies, universities, they all have to participate.
Ajay- What are the top  tips you would give to aspiring technology writers and bloggers (like myself)?

Steve-
1) Learn about non tech subjects, such as history, literature, art and psychology
2) Work on writing clearly for non experts. Avoid jargon.
3) Do reporting
4) Do more reporting

Ajay-The Numerati portrays a math elite which breaks the stereotype of the lonely, nerdy geek. How important do you think is that common people be more educated in math so they are more aware of marketing operations and credit offers?

Steve- I think it’s important for common people, as you call them, to understand basic statistics. More and more of our lives are going to be analyzed and communicated to us statistically. Those who do not understand this will not know to ask the right questions, and will be easily fooled. This is also true within companies. CEOs can be fooled by numbers, just like anyone else.
Ajay- Asia delivers a disproportionate number of science graduates. Yet one generation ago American and European heritage scientists made the trip to the moon with very basic computers. As our lives get increasingly shaped by the Numerati, how important are geo-cultural influences in its membership?

Steve- Most of the Numerati I met in the United States were born outside the U.S. The US has long relied on foreign brains, especially for its technology industry. As the Numerati study people’s lives, the quantitative experts will increasingly need to work closely with linguists, anthropologists, and psychologists. And they’ll need to understand different global cultures and languages. In this sense, the international nature of the Numerati is an advantage.
Ajay- Do you think the shift in money and influence from lawyers and financiers to scientists and mathematicians is temporary or is it here to stay?

Steve-I think it’s here to stay. There will always be lawyers and financiers who make loads of money. But they will have quantitative experts on their teams.

Ajay- What influenced your decision to be associated with Predictive Analytics world?

Steve- I had the privilege of interviewing Eric Siegel as I was researching the book. We’ve kept in touch since then. I think he’s very bright and does excellent work.
Ajay- What does Stephen Baker do when not writing books or articles or observing the world go around him?

Steve- I like to ride bicycles, I like to travel. I love Spanish and French and baseball and music

Biography-

Stephen L. Baker is the author of The Numerati and a senior writer at BusinessWeek, covering technology. Previously he was a Paris correspondent. Baker joined BusinessWeek in March, 1987, as manager of the Mexico City bureau, where he was responsible for covering Mexico and Latin America. He was named Pittsburgh bureau manager in 1992. Before BusinessWeek, Baker was a reporter for the El Paso Herald-Post. Prior to that, he was chief economic reporter for The Daily Journal in Caracas, Venezuela. Baker holds a bachelor’s degree from the University of Wisconsin and a master’s from the Columbia University Graduate School of Journalism.

You can read more about the Numerati at http://thenumerati.net/index.cfm?catID=18 Stephen L Baker is the keynote speaker at Predictive Analytics World and you can check the details here http://www.predictiveanalyticsworld.com/register.php if you want to listen to  him at the event.

You can follow Steve on twitter at http://twitter.com/stevebaker and follow his blog here http://www.businessweek.com/the_thread/blogspotting/

Interview Jeff Bass, Bass Institute (Part 2)

During the 1980’s and early 1990’s, the Bass Institute managed to attract a loyal following with it’s SAS language compiler, ultimately bowing to the financial pressures and technological pressures of the move to the Desktop. In the year 2009, as SAS language gains a new compiler in terms of the WPS, AND computing paradigms begin to shift to cloud computing from the desktop- Jeff Bass, founder of Bass Institute and genius tech coder brings a perspective rich in experience.

If we don’t learn from history, we are condemned to repeat it.

Ajay- Describe your career in science. How would you motivate children in class rooms today to be as excited about science as the moon generation was?

J Bass- My graduate training was in economics and statistics.  I have used that training in ways that I would never have anticipated when I was in graduate school 30 years ago.  But it is still exciting for me.  I started out building microeconomic models, then went on to write statistical language compliers and build health policy macroeconomic models.  These days I develop and articulate health policy to help increase patient’s access to cutting edge medicines.  The company I work for now is very science based and even applies scientific thinking, measurement and testing of alternatives in the business side of its operations.

I spend volunteer time as a guest teacher at local middle schools, high schools and community colleges.  I often talk about math and statistics and have found that one way to help motivate students is to give them “fun” example problems.  I often use an example of the 1969 lunar orbital calculations to motivate basic trigonometry and quite a number of students who say they don’t like math end up loving solving parts of that problem.  I think our school curriculums need to come up with problems and examples that the students find interesting.  I’m not sure our existing curricula processes make this an easy thing to do.  All too often we teach techniques without combining that teaching with strong motivating examples that make learning fun.

Ajay-  What are the changes in paradigms that you have seen across the decades? What are the key insights and summaries that you can provide.

J Bass- Our increasing understanding of biology and DNA is a major paradigm shift that is combining molecular biology and protein chemistry with computer science.  Identifying the human DNA sequence was only the beginning.  Imagine that you were handed the bit sequence of a CD-ROM and were told to figure out what parts of it were a text document, what parts were a JPEG photograph and what parts were an MP3 music file – if you did NOT know the coding schemes of such files.  That’s analogous to where we are today with DNA sequences…we know the ATCG sequence, but we are only scratching the surface of understanding the things that the DNA sequence codes for – proteins, cell metabolism, differentiating cell reproduction. Continue reading “Interview Jeff Bass, Bass Institute (Part 2)”

Jump to JMP: Using Data Analysis in a visual manner

Over the past month or so, I have really begun to appreciate the GUI of JMP. It is very clean and intutively designed. And excellent for a SAS Environment .

Best of all you can easily download a 30 day trial and pricing for this software is quite reasonable.

The worst part of JMP- the droll website. In fact on website, I can deduce something of an Ohri’s Law on Websites.

The better the software, the worse off is the website.

Corollary- The worse off the software, the better is the website in terms of glitz.

JMP is definitely worth a trial for 30 days if you

a) Want to learn a new stats software skill fast

2) Unhappy with visual data analysis of current softwares.

Integrating JMP ‘s functionality with a BI reporting tool is a formidable data decisionmaking tool and it works nicely for me in data analysis I do.

http://www.sas.com/apps/demosdownloads/jmptrial8_PROD__sysdep.jsp?packageID=000503&jmpflag=Y

jmp