強化学習を使って作図問題を解く Chainer Meetup #07, 9th Jun 2018
ఆنͱίϯύεͱChainerRLChainer Meetup #07, 9th Jun 2018[email protected]
View Slide
ChainerRL Ͱ
࡞ਤΛղ͖͍ͨ
࡞ਤ• ఆنͱίϯύε͚ͩΛͬͯతͷਤܗΛඳ͘http://mathworld.wolfram.com/GeometricConstruction.html
σϞ
ͷલʹ
ਤͷݟํֶशϞσϧʹ͢ใʢObservationʣਓؒ༻తͷਤܗར༻Մೳͳ
શମ૾ڥΤʔδΣϯτߦಈ؍ଌ
શମ૾ڥΤʔδΣϯτ[p0_x, p0_y][p1_x, p1_y]……ը૾ͷใ[shape_flag, pi, pj]৽͍͠ਤܗ
ωοτϫʔΫΞʔΩςΫνϟConvMLPMLPConvMLP[p0_x, p0_y][p1_x, p1_y]……ը૾ͷใ[shape_flag, pi, pj]৽͍͠ਤܗ
ωοτϫʔΫΞʔΩςΫνϟ(100, 100)(12, 3)ConvMLPMLPConvMLP(2, 12, 12)[p0_x, p0_y][p1_x, p1_y]……ը૾ͷใ[shape_flag, pi, pj]৽͍͠ਤܗ
ωοτϫʔΫΞʔΩςΫνϟ(100, 100)(12, 3)ConvMLPMLPConvMLP(2, 12, 12)[p0_x, p0_y][p1_x, p1_y]……ը૾ͷใ[shape_flag, pi, pj]৽͍͠ਤܗ= 288
ࢥͬͨ͜ͱͳͲ• ڧԽֶशͬͨ͜ͱͳ͔͚ͬͨͲָ͍͠• ChainerRL ϥΫͰΑ͍• ߦಈۭ͕ؒେ͖͍ͷͰݮΒ͍ͨ͠• AlphaGO ͕ࢀߟʹͳΔ͔ʁ• ίʔυ͖Ε͍ʹͨ͠Βެ։ && ղઆ͠·͢• n ࣍ํఔࣜΛ ChainerRL Ͱղ͚Δ͔ʁ• ՝֎׆ಈ͖ͳਓɺҰॹʹΓ·͠ΐ͏ʂ