timr: The difference between this and normal outsourcing is that I'm not outsourcing the design aspect of it, merely the algorithmic development. Think of it as a netflix challenge style problem. The problem would be specced roughly as:
"Here is an input vector, here is an output vector, write an algorithm that can translate input -> output with 98%+ accuracy". Of course, there would be validation data as well which they never get to see.
scott_s: Of course, lots of research is by necessity vague but a lot of it can be tightly scoped as well. Handwriting recognition, for example, simply involves trying to improve an accuracy measure.
sireat: There's what? 1000 universities in China and I don't read Chinese. Such an approach seems really hit or miss to me.
I understand what you want, but was restraining myself not to say something harsher, because I found your question rather offensive.
What makes you think that a PhD who has spent 20+ years of education becoming an expert in some particular research specialty, is going to want to solve your core business problems on spec, for crap pay and zero equity? If you think this stuff is so easy that you just can contract someone to solve your problems "with 98% accuracy" the way you'd order a graphic design from eLance, then you have an extremely poor understanding of what a PhD is trained to do.
(In fact, somehow I doubt that you'd order a graphic design from eLance, because by now, most entrepreneurs know the value of developing an ongoing relationship with a skilled professional. A PhD is a skilled professional, too.)
timr, I'm not coercing them into anything and I don't have any illusions that I'm going to get a PhD for less than minimum wage. If they don't want to solve my core business problem, that's fine, all they have to do is not bid.
An ongoing relationship is a great thing when it's part of your core business competitiveness. But for some tasks, it's simply a matter of getting something that meets a spec and I don't see why outsourcing is not an acceptable solution for this.
I short, I don't know where your anger is coming from.
"Here is an input vector, here is an output vector, write an algorithm that can translate input -> output with 98%+ accuracy". Of course, there would be validation data as well which they never get to see.
scott_s: Of course, lots of research is by necessity vague but a lot of it can be tightly scoped as well. Handwriting recognition, for example, simply involves trying to improve an accuracy measure.
sireat: There's what? 1000 universities in China and I don't read Chinese. Such an approach seems really hit or miss to me.