Slashdot Mirror

Even if you use UW facilities for research? That's doubtful.

Doubt all you want, the facts are the facts.

Here's WARF:

UW–Madison faculty, staff and students are not obligated to assign their intellectual property to WARF, unless required to do so by federal law or the terms of a sponsored research agreement with a third party.

http://www.warf.org/about-us/f...

Here's the official UW Policy:

he UW is unique among U.S. universities in that it does not claim ownership rights in the intellectual property generated by its faculty, staff, or students, except when required by funding agreements. UW inventors do, however, have an obligation to disclose all inventions created while carrying out university duties, using any university funding, or using university premises, supplies, or equipment. It is the role of the UW–Madison Office of the Vice Chancellor for Research and Graduate Education to perform an equity review for each UW–Madison invention disclosure to determine what obligations may attach to each invention and who may have rights to the invention.

https://research.wisc.edu/proj...

Note that you have to disclose the invention to WARF. However, you don't have to give it to them. You have to disclose it because some (but not all) outside grants require inventions be assigned to the university, and WARF wants to make sure you're following the terms of the grant. However, that is not the University's fault. I speak as someone who has disclosed an invention to WARF. They looked it over, determined that it was not covered by a grant that requires I give up the invention, and told me that I could do what I wanted.

WARF and lawyers are the middleman

By that standard, it doesn't matter who owns the patent: any time there is a patent lawsuit there are middleman because lawyers are involved. Of course, that's a silly standard because they don't "make all the money" as you claimed a few posts ago. WARF doesn't really make money since it's a non-profit, and as noted above, despite your disbelief, you can cut WARF out of the loop.

WARF does pay a substantial share to the inventors: http://www.warf.org/for-uw-inv...

From the link: "20% of royalties (before expenses) will go to you [the inventor]."

Because the Bayh-Dole Act was enacted by a Congress (which funds the NSF) that does not agree with you in the the same sense.

Research money has to come from somewhere. It can come from your taxes, it can come from university money, or it can (and does) come from a combination of the two. Universities can make money by charging students fees or by licensing patents to inventions developed by their staff (and student employees). Money made by licensing patents has the advantage of coming from commercial partners and/or customers using the technology, rather than students working part time jobs and taking out gigantic student loans.

To rip of the AC who also replied to your post, but will never get modded up, "[t]he system becomes self funding and if they do their job right the pool of cash involved should grow exponentially." The AC was referring to the NSF and being wildly optimistic, but institutions like WARF show that the concept is sound -- successful research universities can reduce their reliance on the NSF for funding and/or fund areas that are not NSF priorities by seeking patents related to their research.

Shouldn't research paid for "by the people" be available "to the people"?

It is -- in the same way that a public ampitheater paid for by the people is available "to the people." Not every person has the right to use the stage at any given time, management of the ampitheater may be contracted out to a private operator, and "the people" are more often than not only the consumer of a product embodying the IP rather than a producer attempting implement the IP. Your concern for the needs "of the people" is a touch too abstract.

In addition, the research is available -- that's the very purpose of encouraging patent disclosures -- it is merely the commercialization of the research that is temporarily restricted. Don' like it? Design around it. Or come up with an improvement that is good enough to but you a seat at the table. Both alternate technologies and cross-licensed improvements help, not hurt, technological development.

Seriously? Check out WARF. While technically not part of the government they're really just a branch of the University of Wisconsin. They must have dozens of patents on useful stuff including drugs. Probably the most famous is rBGH.

For universities it's easy: as most of them benefit from public funds, they shouldn't be able to patent anything and release it all under the public domain for the public's benefit.

Well, you have to repeal/amend the Bayh-Dole Act that essentially gave universities the right to patent their findings. I think before that the patents went to the United States government if they funded the research. I know that our friends at the University of Wisconsin (Madison) have courted the government to keep funding them by offering Institutional Patent Agreements. Does WARF sound familiar to you? It should.

There's a lengthy blog post about this that has good quotes and points from both sides including:

Georgia Tech professor Mark Allen said "In a number of circumstances, the competitive advantage afforded through exclusivity [that is, patent monopoly] may be absolutely critical to justify the risk undertaken by a company in developing a product from a promising early-stage university technology, as it was in the case of Cardiomems." Professor Allen, also Chief Technology Officer at Cariomems, did not reveal his compensation from privately-held Cardiomems using the patented technology from his Georgia Tech research.

Susan B. Butts, Dow Chemical Company, had a different perspective: "Although the Bayh-Dole Act has enabled the transfer of technology developed with federal funds from US universities to industry it has also contributed to a contentious climate around the issue of intellectual property (IP) rights which discourages research collaborations between industry and US universities. Second, most foreign universities, which do not have the IP expectations created by Bayh-Dole, allow industry research sponsors to own or control inventions resulting from the research that they fund. This much more favorable treatment of IP is causing companies to do more of their sponsored research abroad [emphasis added]."

And, you know, with how much value we place on intellectual property elsewhere it would seem that the amount of funding and rewards universities are getting for this research is down right laughable. So the Bayh-Dole Act was a very simple solution: let both parties involved benefit from the research and allow the university to reap the benefits of licensing and royalties.

What's a better alternative method for appropriate rewards?

The government also sponsors applied research, often through DoD and NASA. Our major aerospace companies (Lockheed, Raytheon, etc.) receive many millions of dollars in federal research (IR&D) funds each year to work on various projects, most of which are "pre-commercial". That means that the company itself wouldn't make such an investment with its own funds, since it might have a negative hit on its bottom line and therefore affect the stock price. However, we've seen that those research dollars often lead to huge business opportunities for established companies and startups as the results of that work become commercially attractive. One great example of this is the Internet, which comes directly from DoD ARPA funding.

Of course, government agencies don't always make the right call about what ideas are worthy of investment and who should get the money. In some cases, the proposals are poorly developed or the reviewers give them low ratings. It's also possible for political considerations to affect the decisions of who and what gets funded, particularly if the science advisors don't have strong science credentials. But it often works out pretty well: polio vaccines, man on the moon, jet airplanes, and so on.

I didn't have the H1-B debate in mind in my comment about talented immigrants. Let me explain what I meant. Suppose that a top engineering graduate of Indian Institute of Technology (or some other outstanding foreign institution) wants to work for a Ph.D. in computer science or engineering at MIT, Carnegie Mellon, or some other top US school. First of all, the US has become much less attractive to foreign students as we have made it more difficult for them to get visas; Canada, the UK, and Australia are attracting many more such students than they did five years ago. Second, it used to be much easier for the Ph.D.s to stay in the US after they finished their degrees. When you look at the top management of Silicon Valley companies, you can see that many of the founders and executives were not born in the US. (Sergey Brin and Andy Grove are just two examples out of thousands.)

I think that keeping these talented people and their entrepreneurial energy in the US is a good thing. I doubt if they take jobs from Americans; indeed, the opposite is true: they create companies that employ thousands of people and generate wealth. The current policy, though, requires most of these people with advanced degrees to return to their home country soon after they have completed their studies. Not only do we lose them to our own economy, but we set them up to compete against us. When we buy Acer and Lenovo computers, we are sending money to Taiwan and China. As long as the US isn't doing a very good job of attracting our own students to careers in math and science, we have to import that talent in much the same way that US sports teams now scour the world for the best talent.

Apologies for the long message, but there's a lot to be said on these issues, and they are all strongly connected to science policy.

Inventors' Share
The inventors receive 20 percent of the gross royalty revenue generated by a licensed invention.

Laboratory Share of the Annual Grant
Of the first $100,000 generated by each license agreement, the inventors' laboratories receive a grant equaling 70 percent of the gross royalties. For example, if an agreement generated $50,000 in royalty revenue over its lifetime, the inventors' laboratories would receive 70 percent of $50,000, or $35,000.

Graduate School Share
After the laboratory and department shares have been allocated, the remainder of WARF's annual grant is given to the UW-Madison Graduate School. The Graduate School uses this money to support a variety of projects and programs each year, including:

        * The Graduate School Research Competition
        * The Romnes Early Career Awards and the Kellett Mid-Career Awards
        * Named professorships and graduate fellowships
        * Campus building projects

Most of the money of goes back to research? You don't know that.

Don't forget the IP lawyers the universities have. You know, the ones that 'protect' all those patents and copyrights that the Universities own? I know the WARF is not a body to be reckoned with.

But then again, the people being prosecuted here are the students. And why the hell would anyone like WARF want to waste resources on a handful of students when they've got other things to do?

The University of Wisconsin has a foundation (The Wisconsin Alumni Research Foundation - or WARF) that takes care of patenting research and then allocating the profits. Yes, the University of Wisconsin is state funded, but that doesn't cover the entire cost of a student's eduation. The profits of patents go back in to the University. The professor will not get all of the profits. The majority will go back into the school to help prevent major tuition raises.

It's a pretty good system (the first of it's kind) and a lot of other schools have set up similar systems.

For more information, check out the WARF website: http://www.warf.org/

Well, it's quite obviously because a Support Vector Machine is inherently linear, and to make it nonlinear, you must insert a nonlinear kernel which you need to select by hand.

Not true.

"This invention provides a selection technique, which makes use of a fast Newton method, to produce a reduced set of input features for linear SVM classifiers or a reduced set of kernel functions for non-linear SVM classifiers."