I precieved that, the rights to use the algorithem in any way, will be granted to the sponser, but not transfered. It means that the participant team also may use them in scientific publications or commercial products as it is not denied explicitly. Is this right?

 

Yes that's my understanding - you're providing a license to your algorithm, but it doesn't mean you can't use it yourself. Note however that you need permission to use the data sets for other purposes.

"Algorithm" is not clear. I have two kinds of algorithms, one is the prediction formula and is the final product of the analytics, and the other is the data mining tool.

Guess that Herritage refer to the forst one...

 

I have exactely the same question concerning both:

- The formula

- The tool to get the formula

 

Another question is: can we use a commercial tool (of course with copyrights) to get the formula?

 

Thanks

 

This is also a neat but very important differentiation that I am curious to know about.

Rule 21 specifically says you need to provide a license to "the algorithm used to produce the entry".

That doesn't rule out using a commercial tool to implement that algorithm. However, if you use an algorithm embedded in the commercial tool, and that algorithm is either not documented or not licensable, then you won't be able to comply with the terms of the agreement.

Winners of other Kaggle comps have used commercial tools such as Matlab. They have used these tools to build their own algorithms, or used Matlab implementations of algorithms that are fully documented in the literature and not restricted by patents of other IP issues. There is no problem with doing that in this competition.

About item #21,
there's a common confusion, I see it often at my work, between a prediction algorithm and the analytics software that produces it, which is also called mistakenly algorithm. In the RULES document the term algorithm is used interchangeably 16 times!

In principle, it is right that we give all of our analytics results from this contest, including the prediction formula. But, it's no one business to demand any rights over the software that produces the results. All that can and should be asked by Heritage is that the software is lawfully used, as indeed is requested in t#20.

I think that if the winner uses a commercial software to extract a model, then he will have to provide both the model and a licence to use the commercial software. But with at least 30.000$, it should not be a problem to buy a licence... Moreover, since it is a very good publicity for the commercial software, I am sure there is a way to get it for free...

They've posted a clarification... but unfortuately it doesn't clarify anything. What is meant by "the algorithm"? If I use an open-source, copy-left restricted machine learning package, am I disqualified because I don't have rights to grant non-copy-left exclusive use of the package? Taken to an extreme, it would rule out using anything but "bare metal" and even that has proprietary microcode. Obviously something "in between" is meant, but what?

Was the fact that it was an exclusive license grant part of the earlier version of rule 21? I don't remember that when I read it yesterday, but I might have been losing concentration by that far into the page.

The license quickly went from non-exclusive to exclusive use.  The caveat being business use of the entrant.  But internal business use of the entrant could be anything that doesn't involve transferring or granting a license to a third party.

For clarification:

Do you intend that I be able to create a derivative work and sell the service to others for a profit?   Is this considered an internal business use? 

Can I continue to develop and use the improved software after the contest ends or would the exclusive license prevent me from doing that and using it to deliver new and innovative tools for profit to the market?

Can Kaggle create and sell a program based or provide a service for profit basded on this work?

I would suggest that having an Open Source License would make more sense for this contest.  That would allow anyone to use or improve the prediction software as long as they contribute the improvements back to the community that is using/developing it.

An exclusive license that prevents me from ever writing this type of software again without violating the terms of your agreement even if there is no payment of any prise is quite extreme. 

@shaunc: I share your concern. I'll try and clarify this with HPN. I'm sure we'll find a wording which satisfies everyone.

ItsMeRonT: Just to clarify, the license is not to Kaggle, it's to HPN. The license doesn't give Kaggle any rights to use the algorithm.

OK just got off the phone to HPN. They definitely don't intend for the language to result in anyone being restricted either in terms of the tools they can use (as long as it doesn't stop HPN from utilizing the model), and will be issuing a clarification in the next couple of days regarding the intention of this clause.

I made an early copy of all of the competition documents:

The primary state of Rule 21 was:

"By registering for the Competition, each Entrant (a) grants to Sponsor and its designees a worldwide, non-exclusive, sub-licensable (through multiple tiers), transferable, fully paid-up, royalty-free, perpetual, irrevocable right to use, not use, reproduce, distribute (through multiple tiers), create derivative works of, publicly perform, publicly display, digitally perform, make, have made, sell, offer for sale and import the entry and the algorithm used to produce the entry, as well as any other algorithm, data or other information whatsoever developed or produced at any time using the data provided to Entrant in this Competition, in any media now known or hereafter developed, for any purpose whatsoever, commercial or otherwise, without further approval by or payment to Entrant (the "License") and (b) represents that he/she/it has the unrestricted right to grant the License."

 

Later, it changed to:

"By registering for the Competition, each Entrant (a) grants to Sponsor and its designees a worldwide, exclusive (except with respect to Entrant), sub-licensable (through multiple tiers), transferable, fully paid-up, royalty-free, perpetual, irrevocable right to use, not use, reproduce, distribute (through multiple tiers), create derivative works of, publicly perform, publicly display, digitally perform, make, have made, sell, offer for sale and import the entry and the algorithm used to produce the entry, as well as any other algorithm, data or other information whatsoever developed or produced at any time using the data provided to Entrant in this Competition (collectively, the "Licensed Materials"), in any media now known or hereafter developed, for any purpose whatsoever, commercial or otherwise, without further approval by or payment to Entrant (the "License") and (b) represents that he/she/it has the unrestricted right to grant the License. Entrant understands and agrees that the License is exclusive except with respect to Entrant: Entrant may use the Licensed Materials solely for his/her/its own patient management and other internal business purposes but may not grant or otherwise transfer to any third party any rights to or interests in the Licensed Materials whatsoever."

 

I think that the algorithm/ Engine/ Software/ Ideas whether proprietary, free or open, in-house or commercial, published or unpublished, that are used for creating of the predictive algorithm can not be included neither it sure violates others' IP rights in some way. But the resulted predictive algorithm may be shared in an exclusive manner.

 

But some questions raise:

1- What happens to those made aggreement before this change?

2- What happens if a Entrant is an organziation which is already provides very widespread services to patients, among the country? (e.g. governmental, NGO sectors)

3- What if the Entrant is a service provider to healthcare organizations already with software products in use?