So I now have my paper on methods for one time programs on ArXiv and thus am free to write about the project and my experience filing for patents on the idea.
Details of the patent can be found here.
This review is to remind myself of the steps taken in filing for a patent (and other patents which were started and not completed). It may also serve others who are considering patenting some of their academic research to consider some of the possible difficulties that they may encounter.
This may also serve as a guide to anyone looking to form intellectual property as part of a commercial venture - especially if they are doing this before getting venture capital or other funding (VC's are notorious for not agreeing to non disclosure agreements so it may be a risk to go into details on a technical pitch - incase they kick back the idea to a business they already have equity in).
That patent describes that quantum resources may be used to implement on time programs, generally.
In it we also discuss error correction, and specific encodings, as well as a very specific setup using photons.
A layman description of the idea is described here.
To Patent or not to Patent
You may think that of course you should always patent right! Wrong. Apart from the clear cost and reward of owning the patent there are other things to take into account... This is true even if you arn't paying for the patent yourself.
The decision to patent the method for implementing one time programs using quantum resources was taken quite early in the project. This proved to be both a hindrance and a blessing.
The boon of applying for a patent early in the research timeline is that it can be granted by the time you finish university. The timeline appears to be around 2 years from the initial technical disclosure to being granted. This may involve review stages; a process similar to fighting with referees when submitting a journal publication. There is a big difference between patent filled and patent approved, as there is with journal submission. In short anyone can file a patent given enough money and the inclination, whether it awarded is a different matter.
Unfortunately, like journals, your referees may not always be totally in sync with what you are showing (especially with all of the legalese language used). This happened to me and a rebuttal stage was necessary after the first round as the patent attorney found some papers showing quite different results, but believed them to invalidate our claim. They didn't; they showed dramatically different results as we later explained, nevertheless this is something that shouldn't worry you, but something you may need to take into consideration.
The major disadvantage with filling early in a research project is that, until the patent is filled, you cannot publically disclose details to the public.
Now for me this was a huge issue. In research, and especially I believe in theory, collaboration and discussion of ideas dramatically increase productivity. When you have an idea and are not sure how to make it work, a quick discussion with another physicist down the hall who had previously worked in the area can save months of literature review and trail and error. It is very possible that people you talk to will know of past results implying a particular method is likely to result a dead end, of if just the mathematical concoction you are looking for already exists in some obscure math journal. On top of that, its just a much more pleasant work environment to have a little human contact in the day.
However - if the piece of maths you are trying to derive is related to the patent you are in process of filling, then thing become difficult. Suddenly you are not able to describe the problem you are trying to solve in detail, which ultimately undermines the workflow of research. No longer can you post on stackoverflow, or grab a coffee and puzzle with friends by a whiteboard. You certainly can't give a research talk or present a poster at a conference.
Now, if you are the only person working on this, as I was, then PhD life immediately becomes a lot harder and more isolated. You really are alone on the problem, and you cannot get help until the problem is done.
For me this has been a huge issue. I attempted for approximately two years to form an analytical bound describing the distinguishability of error corrected one time programs using quantum states (if that sounds complicated to you... that's because it was...) without much success. I believe if I had been able to discuss the work it would have only taken six months or so before the problem was changed to allow me to find a numerical solution (which is what we eventually did, and something that could have been done in a few months). Although the paper is now out, this time sink has affected the number of publications I have in my PhD, and my ability to travel to conferences and network in the community. It also made for a rather unenviable two years of isolated work banging my head into a brick wall.
So that's the major opportunity cost to bear in mind.
For the same reasons I've decided not to file on a two other patents despite doing all the work through to technical disclosures; one concerning the cryptclip, which I was asked to file for by the university after showing a prototype, and another idea (relating to classicla one time programs). I will be first to admit that there is dubious reason to file the patent if I do not intent to develope the ideas afterwards other than having CV filler. However, the lack of ability to communicate ideas was the primary motivator not to patent, especially as I come to the end of my PhD.
The first step is to create a technical disclosure. For me, the format of this was providing by the graduate office at my university. This is simply a run though of the method, including the details of the inventors and the date on invention.
Now, if you are running a collaborative project then all the associated universities may want to file a patent through them.
For universities patents aren't just valuable in terms of generating revenue in the future, but as a key performance indicator of research at the university. For SUTD, a new university, it's a great thing to be able to tell government and industry sponsors the number of patents generated when fishing for future investment or demonstrating the quality of your leadership. There may also be some other incentives to the tune of "if we don't spend this money then we wont get more!" and "we need more work for this department or else I'm less important as its head", both of which is incredibly common in academia. These both would mean there is a readiness to spend money and get lawyers involved. This is slightly worrying as it does mean there is something of a frenzy to patent as much as possible, regardless of value, but the interests also align with those of student and Professors, who think they look great on a CV, and lawyers, who are happy to take the payments to file a patent. Luckily this isn't really abused as people only seem to tend to go to the university intellectual property staff if they do have a good idea, but it certainly could get out of hand should some patents become more important than papers. Its worth noting that you can add patents to both your google scholar profile and LinkIn; so this could potentially happen.
In our case, we actually filled with the University of Vienna, who I believe paid the lion's share of the lawyer fees, with SUTD working in conjunction, just to add complication. This meant that we worked directly with the patent office in Austria.
Inventors and Ownership
Now, although my collaborators and I are the inventors of the patent, we do not (technically) own the rights to the patent. The company, or this case university, who pays for the filing of the patent, owns the patent itself.
However, universities often want to see a return on their investment in obtaining the patent and so want someone to develop the technology into a company. SUTD, like many other universities, will allow and entrepreneurial spin off coming from the university to have free rights to use the patents for some number of years. This is a win-win for the university - they now have a provably valuable patent - which they may license to some other company to use, and they university proves it can spin out companies.
Dealing with Lawyers
Unfortunately, after submitting a technical disclosure, and meeting with the hired lawyers who will be submitting the patent claim, a document is formed to try and explain the technology in a horribly contrive manner. Legalese.
This is quite possibly worse than Singlish.
What was a very understandable description becomes nigh indecipherable (especially something describing a quantum invention…). They would make great cryptographers. Certain phrases such are "State of the art", "Person skilled in the area", and "Novelty" have very specific meanings though that you need to be careful with. During the process of translating the method to a legalese text, errors are bound to occur, so this document will have to be parsed several times to make sure all the technical details are still accurate.
The final document should, as well as forming the most general notions possible to cover as much as possible for the future, as well as your specific implementation, have a list of claims.
A person attacking your claim would need to invalidate all of the claims to invalidate your patent completely, which go from most broad to most specific.
This entire process takes approximately a year - or at least did for me.
After this the patent is sent to an attorney at the office who will try to understand what the hell you are talking about, and try to find documents to invalidate it. Some documents you will need only to reference, again in a similar manner to a journal article, but some may invalidate the claims (mostly on the impact of novelty) in which case you will have to either rebut them or drop that specific claim and any other claims that cite that claim.
Throughout this process its worth bearing in mind that the lawyers, despite working very quickly in my ecperience are not working to the same time pressures you and your collaborators are. They will not be paniced by conference date or teaching time schedules but only by filling date and their other work for different clients.
PCT and world patent differences
Now - just because you have a patent filed doesn't mean you have the right to use it in all countries. The price to use the patent increases ever year, and lapses completely after 20 years. The increase and starting price is also different ins different world regions.
To initially get worldwide coverage, the patent can be filled as under PCT, which is what we did, before deciding which regions are most important at some later date.
Ok - that's all for now. I may write some further updates on the experinience soon.