More spitballing about "Occam's Climate Model Prize" (catchier I think):
Make the resource limitation of the competition classes be powers of 10 dollars spent to run the verification in a standard cloud service, but start with 0$ on Google Colab within the free resource limits of that service so anyone can verify and anyone can develop a model at that level of resource consumption. Competition-classes $1, $10, $100, $1000, $10,000, $100,000, $1,000,000... each with correspondingly larger corpora of climate data. How valuable is a good climate model? Aside from taking advantage of the primary function of money as the standard metric for resource allocation, there is a synergistic subtext here: $ spent corresponds (roughly) to CO2 emission from energy consumption by the cloud service's computation. Another benefit is that prize entries could be required to pay f($10^class) per submission where f(1)=c>1 and f(x) is identity-asymptotic as x->inf. This covers the cost of verification by the judging committee, and discourages frivolous entries. The exact formula for that would be pretty close to $10^class, with greater overhead for the lower values of <class>. There would be a controversy over the choice of data at each class, but that controversy would illuminate the salient issues. On Thu, Jul 2, 2020 at 2:17 PM James Bowery <jabow...@gmail.com> wrote: > Just spitballing here: > > Occam's Razor Models of Climate Change > > The competition classes would be defined in terms of the scale of the > datasets available as well as the computer resources. > > > > On Thu, Jul 2, 2020 at 2:06 PM Ben Goertzel <b...@goertzel.org> wrote: > >> X-Prize tends to be driven by corporate donors who fund prizes, so if >> you/we could convince someone to fund a prize purse for " a series of >> AI competitions based on resource constraint classes." then X-Prize >> Foundation would likely strongly consider it... I do know the X-Prize >> folks fairly well ... >> >> If we coupled " a series of AI competitions based on resource >> constraint classes." with AI explainability somehow it would be buzzy >> and maybe make it easier to get sponsorship... >> >> But corporate sponsors won't like abstraction so much, we'd need a >> wizzy idea for specifically what the competition should focus on... >> >> ben >> >> On Thu, Jul 2, 2020 at 12:02 PM James Bowery <jabow...@gmail.com> wrote: >> > >> > On Thu, Jul 2, 2020 at 1:23 PM Ben Goertzel <b...@goertzel.org> wrote: >> >> > Can we agree that, regardless of the frontier search heuristics, it >> would benefit AI, both general and narrow, to wave about the garlic of >> "Resource Constraint" providing "competition classes" _within_ which >> metrics (of whatever justification) are fairly compared? >> >> >> >> >> >> Yeah, while I feel it's valuable to have folks pushing on "what >> >> capabilities can we achieve with all the resources we can muster", I >> >> also agree w/ you that it's valuable to have resources and attention >> >> focused on "what capabilities can be achieved with so-and-such variety >> >> of limited resources" >> >> >> >> I'm not sure what practical steps you are suggesting for whom to take >> tho? >> > >> > >> > Excellent question and indeed similar to one I addressed when >> approaching various private sector fusion energy initiatives leading to >> circa 1992 legislative language to privatize the fusion program in a series >> of ~10 objective milestones each awarded a $100M prize, and in approaching >> various amateur rocketry groups to establish a small prize for achieving >> 100km altitude that led to a much larger $250k prize for the same >> criterion, thence, one could easily imagine, the Ansai's stepping up to the >> plat for the X-Prize Foundation for a manned version. >> > >> > Which brings to mind Singularity University's executive founder Peter >> Diamandis... >> > >> > It seems to me Kaggle is hopelessly mired in unprincipled approaches to >> machine learning competitions, but there may be hope for the X-Prize >> Foundation to establish a more principled approach to a series of AI >> competitions based on resource constraint classes. >> > >> > >> > Artificial General Intelligence List / AGI / see discussions + >> participants + delivery options Permalink >> >> -- >> Ben Goertzel, PhD >> http://goertzel.org >> >> “The only people for me are the mad ones, the ones who are mad to >> live, mad to talk, mad to be saved, desirous of everything at the same >> time, the ones who never yawn or say a commonplace thing, but burn, >> burn, burn like fabulous yellow roman candles exploding like spiders >> across the stars.” -- Jack Kerouac ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T100f708e32ae7327-Md2f1aacf3f0f0083b29d84d5 Delivery options: https://agi.topicbox.com/groups/agi/subscription
