tag:blogger.com,1999:blog-5880610.post8321688818471182893..comments2024-01-13T18:57:18.243-05:00Comments on Information Processing: The Mystery of GoSteve Hsuhttp://www.blogger.com/profile/02428333897272913660noreply@blogger.comBlogger6125tag:blogger.com,1999:blog-5880610.post-28763669190900858062014-06-03T22:05:14.366-04:002014-06-03T22:05:14.366-04:00Computers made massive progress in the last decade...Computers made massive progress in the last decade. They used to be absolutely terrible, now most amateurs can't hope to beat them. I think 10 more years should do the trick, but it's hard to predict since the budget allocated to Go AI research is (understanbly) very small.alsnoreply@blogger.comtag:blogger.com,1999:blog-5880610.post-18306178626663000802014-06-02T05:04:59.331-04:002014-06-02T05:04:59.331-04:00I'm referring to a specific algorithm that peo...I'm referring to a specific algorithm that people have applied to other games. It's not just a combination of buzzwords. <br /><br />This approach would not achieve state-of-the-art results in playing Go, but I'm interested in seeing what kinds of strategies it would develop and where it would fail.Alex Lambnoreply@blogger.comtag:blogger.com,1999:blog-5880610.post-65098958049181507062014-05-27T15:54:15.027-04:002014-05-27T15:54:15.027-04:00I'm sure people have tried lots of buzzwords a...I'm sure people have tried lots of buzzwords and will continue to do so.5371noreply@blogger.comtag:blogger.com,1999:blog-5880610.post-34074576196970753812014-05-27T05:51:47.538-04:002014-05-27T05:51:47.538-04:00Decapitation strategy gives rise to Chess. Contain...Decapitation strategy gives rise to Chess. Containment strategy gives rise to Go.<br />However, on a spherical surface, active containment of other strong countries is equivalent to itself being contained.<br />It is interesting that Russia might be able to exert containment in the third direction by withdrawing support to the Space Station and maintenance of the GPS satellites.<br /><br />dxie48noreply@blogger.comtag:blogger.com,1999:blog-5880610.post-87062939056014235452014-05-27T04:33:06.446-04:002014-05-27T04:33:06.446-04:00Has anyone tried deep neural networks + reinforcem...Has anyone tried deep neural networks + reinforcement learning for playing Go?Alex Lambnoreply@blogger.comtag:blogger.com,1999:blog-5880610.post-65220716391414555132014-05-26T15:48:10.913-04:002014-05-26T15:48:10.913-04:00It probably won't surprise anyone to learn tha...It probably won't surprise anyone to learn that my own prediction is - never.5371noreply@blogger.com