Can organizers please remind us how new IP is handled in the Challenge?
At the moment we have on the Leaderboard around 5 poorly scoring solutions and one that scores well. Suppose this solution (or my teams eventual solution) involves some novelty, as opposed to say just regressing images against scores using an MNIST variant. That is, suppose the solution involves some original thinking and not simply an application of some well known and thus fairly public domain technique. In short, suppose that the solution is patentable and commercially valuable.
The solution will necessarily be derived in part by observation of data supplied by a number of other institutions. Some of those institutions may have busy, well staffed and aggressive patent departments, and be highly attuned to commercialization.
So in the context of all that, what is the scenario in terms of ownership, financial participation, patenting and commercial development, for the winning solution, when the Challenge terminates?
Created by Lars Ericson lars.ericson I'm actually very pro open source public domain. So I'm looking for the biotech equivalent of GNU or MIT open source licenses that put the idea and it's application solidly in the public domain. Also that nothing in the setup provided by the organizers or their goals in sponsoring the challenge inhibits open sourcing. Just kind of double checking on that score. Because I don't have the resources or the patience or the mindset to close-source, so I am commiting to Door #2 by default. Hi Lars,
For sure. I get that these are important points!
However, I don't think we can provide concrete answers to your questions. Someone in Europe or Asia is certainly operating under different laws than someone in the US. The general advice that's been given in past Challenges is: if you require a patent, it should be done sooner rather than later (e.g. https://www.synapse.org/#!Synapse:syn8404040/wiki/478424). I am somewhat familiar with the process of patenting chemical entities in the United States - my understanding and experience - *with the reminder and caveat that I am not a lawyer* - is that in the case of drug candidates, [public disclosure eliminates the ability to patent an idea](https://www.wikiwand.com/en/Prior_art). I imagine that interpretations vary by locality, subject area, and many other things that I'm not aware of!
My overarching advice remains the same - if this is important to anyone, they should consult a specialist in this subject area. Hi @allawayr, although my Dad and my uncle and my grandfather were all patent attorneys, it didn't rub off on me. However I am guessing that the top solver on the Leaderboard would be happy to build a business around the solution to this problem. My erstwhile PhD thesis advisor from back in the day is now heavily in biotech. He works at NYU and NYU has a department full of attorneys that files patents on everything he works on. So I'm asking the question in the spirit of that aspect of the biotech culture.
To render the point moot, I will place my brilliant inspiration in the public domain with this post: The idea is fairly simple. Use a public domain 3D Python computer animation package to build a visual model of hand and foot skeletons. Parameterise that model at each joint to visually reproduce narrowing and erosion features at each joint. Suppose there are 86 joint parameters, plus a few for width and height and narrowness and skewness of the hand or foot. Scale these model control parameters so that they are they same as the SvH method.
That is, if you put in a 3 for knuckle erosion, then a radiologist looking at the modelled knuckle would score it a 3. Project the 3D image to the plane. Take the difference of the image and an X-ray. Use Levenberg Marquardt to vary the parameters of the model until the distance between the model and the image is minimized. Then read out those parameters as the joint scoring. Dead simple. That's my patent application. I just have to implement it. That said, having expressed it, will somebody patent it out from under me and tell me I shouldn't do it? Should I patent the idea to prevent that? Will expressing the idea in this challenge protect me or the idea? Will participation increase my liability in some way?
Just curious. I'm asking a fairly vanilla bread and butter question which needs to be pretty clear for people who want to participate and promote doing applicable science in this kind of challenge framework. Hi Lars,
Here are what the rules state on IP Rights:
>10 . INTELLECTUAL PROPERTY RIGHTS
>a. DREAM/Sage make no claim to ownership of your Entry or any intellectual property that it may contain.
>b. In the spirit of reproducibility and open science, in order for a final Entry to be deemed as a Challenge Best Performing Entry and for the corresponding Team to be declared a best performer and to qualify for Challenge prizes, the Entry may not include proprietary content such as proprietary software or databases (including, but not restricted to, software/resources that require usage licenses and agreements) in their source code files. The Challenge organizers may inspect and confirm that all winning final Entries abide by these stipulations.
>c. When submitting source code, participants should provide it to Sage/DREAM under an open source license of their choice. The license must permit DREAM/Sage to distribute the code to the public for research and development use via Synapse, GitHub or a similar code repository. As between DREAM/Sage and participants, participants retain copyright to their submitted code. If participants do not provide information on licensing, their submitted code shall be deemed submitted under the FreeBSD license (http://www.linfo.org/bsdlicense.html) and by submitting without such information Participants agree to, and represent that they have all rights necessary to allow, such submission under the FreeBSD license.
>d. Those Challenge Participants who do not wish to submit source code or workflows may, if feasible for the Challenge, instead submit predictive algorithms for scoring. But without submitting source code under an open source license, Challenge Participants will not qualify as a top performer or receive any Challenge Prizes.
>e. You are free to discuss your Entry and the ideas or technologies that it contains with other parties and you are free to contract with any third parties as long as you do not sign any agreement or undertake any obligations that conflict with any agreement that you have entered into or do enter into with DREAM/Sage regarding your Entry. For the purpose of clarity, you acknowledge that the intent of the Challenge is to encourage people to collaborate and share their ideas and innovations.
>f. By submitting an Entry, you grant DREAM/Sage certain limited rights as set forth in these Official Rules. By submitting an Entry, each individual Entrant and each individual member of a Team Entrant grants to DREAM/Sage the right to review your Entry and to have the Challenge Organizers and their designees review your Entry.
>g. A primary objective of DREAM Challenges is to stimulate ongoing collaborative research for biomedical impact. To further this objective, you grant DREAM/Sage the right to make public and publish your qualified final Entry(s) after the results of the Challenge are announced.
Not being a patent lawyer, I can't definitively speak to how providing the method and code in an open-source manner (item 10c) or other parts of this clause might affect the future patentability of a method. If this is of significant concern to you, you'd probably need to consult with one!