Category Archives: breakthrough

AI 2017 part 2 highlights/ trends

hi all. 2nd half AI in 2017 was very fast paced and as usual its hard to keep up. many headlines on the subj. this is a 2nd half review.

seems like some of the biggest news is the increasing dominance of China in the area after the govt announced major initiatives, funding, and resolve/ determination.[b]

google/ deepmind continues to dominate headlines.[a] the other big news is googles alphago and alphazero that no longer required human training games![a2] officially its under the name of “reinforcement learning” but my feeling is that this technology is something like “directed learning” where the AI is manipulating its environment to “detect/ seek/ digest novelty,” and that this will be a key, paradigm shifting trend in the near future. the same algorithm also works in chess/ go!

starcraft[a3] and Dota are starting to show up in AI engines/ research and OpenAI just announced breakthru championship level play vs humans in Dota, and Deepmind is attacking starcraft. despite all this major domination, it looks like games will not move out of cutting edge research anytime soon.[c4]

another notable shift/ trend is that AI/ machine learning is starting to couple stronger with physics and robotics. eg Deepmind humanoid walking simulations etc.

Continue reading

Advertisements

secret/ blueprint/ path to AGI: novelty detection/ seeking

hi all. kurzweil wrote in 2006 “the singularity is near”. foreboding words! but today, still maybe more of a feeling than a fact. definitely, the AI field has started to mature into a new steady advance period/ era in the last few years, also with a burst of energy/ enthusiasm/ innovation heralded by the Google Deepmind acquisition in 2014, and other massive shifts toward increased investment by large corporations and govts. the Musk Open AI initiative was announced in 2015.

the other massive milestone is the ready conquering of Go by Google in 2016 by AlphaGo. in late 2017, a new version AlphaZero was announced that plays superior to AlphaGo (at “beyond human grandmaster level”) after learning merely from the rules and reinforcement learning, ie no example human-level play presented as training whatsoever. AlphaZero also plays grandmaster level chess after learning “from scratch”. this breakthrough is not fully/ widely appreciated in some ways. it is the first case of a potentially more general algorithm for AI emerging from “previously relatively narrow” study of AI in games.

AI has the terminology “weak AI” and “strong AI” for different levels/ sophistication/ “ability”. more recently the term AGI, Artificial General Intelligence has been coined.

Continue reading

norbert blum P vs NP attack goes CYBER-SCIENCE-VIRAL!

⭐ 💡 ❗ 😀 😎 ❤ hi all! want to write a lot on this, have all kinds of impressions and things to say on one of my favorite topics in the world for 2½ decades now and which this blog is partly dedicated to, but am starting out by just collecting some links vacuumed up mostly today for others to check out, the top locations for commentary. unf there is not really a key/ central place so far where all the action is happening. will be adding to this post as time goes on. its very hard to figure out when to time a post wrt very fast moving developments like this, but my trigger just hit mostly right after RJLipton[a10] blogged about it. oh yeah, that comment by gowers that “I think it may reach the level where the experts feel that it needs to be checked” is highly triggering for me too (referring ofc to himself in 3rd person, maybe being more than a bit )( coy about this— doncha just love hardcore mathematicians! which reminds me of the joke about shoelaces…) … oh and then theres also fortnow who literally wrote the book on it! (oh yeah wrt that theres this by RJL too!) 😛

it looks like the Norbert Blum paper[a14][a15] basically went “pop/ scientific viral” today or so probably after mentions on Reddit[a8][a9] and Hacker News.[a11] of the experts Trevisan[a1] was the first to wager a reaction along with some curious onlookers like Baez[a2] and verging-on-jester-or-gadfly-or-even-troll Motl[a5] who along with some mild conspiracy theory about CS research(ers) (lol!) says, wrt the #1 top problem in TCS…

Except that unlike the case of string theory, there exists absolutely no rational evidence that there exists something stunning waiting to be discovered.

@#$% 😡 👿 ❗ whoa, fighting words and/ or near-sacrilege there! excuse me, what is “stunning” is that the problem with $1M prize over 1½ decade has eluded/ stymied/ thwarted the worlds greatest scientific minds for almost ½ century now! and the proof either way, in whatever shape or form, will be assuredly stunning to anyone who isnt jaded or a nihilist! hey physics bozo! think of a great experiment that guaranteed leads to an extraordinary answer either way! arent there any examples like that in your field anyway? or is maybe someone looking a little envious over there? or out of the loop or maybe dont even know how to flirt at all? :mrgreen:

the TCS stackexchange post by a n00b newbie (who didnt know enough to not ask! hows that for “zen beginners mind”?) is now up to 136-4 votes and 28K hits (thursday eve 8/17).[a3] (joy that the lame heavyhanded/ constricting/ dictatorial policy of the self-appointed cyber spoilsports and partypoopers also long criticized here in this blog, is overruled by striking mass opposition!) that post slipped my radar for a few days! found the paper on the 1st day (monday this wk) but thought it might stay “unviral” (you know, like those killer viruses locked in the… thawing… arctic, right?) based esp on the last Hauptmann attempt.[a12][a13]

discovered the big commotion today via reddit myself and spent excited hours poking through sites! but my hits on the Hauptmann blog post had been creeping/ climbing up massively all week, merely from google searches (many from germany), and figured it was due to questions on the Blum attempt. turns out theyre both profs at the same university! (Bonn/ Germany) pardon me but wtf?!?… some kind of story there eh? are any reporters listening?

cyberspace is such a joy sometimes, in my quasi-bipolar days, mostly-sometimes seemingly very flat for days, weeks, months, even years at a time, this is a massive manic spike for me… the adrenalin is flowing! didnt sleep as well this week! there is so much cybersynchronicity to this, must admit, its a very rare event, am myself (and defensively, some others are too) having a bit of a mini-cyber-orgasm so far… 💡 😮 o_O 🙄 ❗

Continue reading

qm computing summer 2017 update

hi all, seem to have reached a momentary lull in my own schedule and also the QM computing news that typically runneth over. maybe a moment of breath to clear out the queue. years ago it was cutting edge to talk about it but now it looks like a lot of the media is attuned and into covering the race[k] esp with inroads by big corporations (Google, IBM, Microsoft, Intel) into the area. Intel seems a relative newcomer but did finally show up on the radar.[a10]

can recall some of the earliest QM computing books came out around late 1990s eg 1997 or so, so now its been almost exactly 2 decades. here is a technology with immense promise but also immense engineering challenge.

Continue reading

Google Deepmind strikes again: AlphaGo 3-0 over worlds best player Ke Jie

hi all Google Deepmind has pulled another rabbit out of the hat and won 3-0 against the worlds reigning Go champion Ke Jie who coincidentally is also Chinese, only about 1yr since last historic match.[a] the 19year old was reportedly literally in tears.[a17] quite a different reaction than kasparov ~2 decades ago! looks like emotions really run the gamut eh? last years match with Lee Sedol had him apologizing for/ to humanity at the end! reminds me of the so-called “5 stages of grief” concept from psychology… where the “loss” is one of mankinds most priceless/ treasured possessions, its uniqueness wrt intelligence/ mind! o_O

theres a brand new AlphaGo movie/ documentary on the prior history, cant wait to see it, has anyone else?[a8] what do you think? how does google/ deepmind top this? apparently they cant and actually theyve announced theyve decided to avoid any further human media/ spectacle matches! or is that just my interpretation? Hassibis is quoted as saying AlphaGo will stop/ retire, but not sure exactly what he means by that, its a big question mark, and the funky/ choppy chinese translation certainly isnt helping any.[a17] doesnt it make some sense to continue to refine it some, or release parts or all of it, do some maintenance, etc? it seems even after it beats the top human there is some small )( suspense left. (have always thought it a bit of a historical loss verging on tragedy that IBM Deep Blue chess hardware/ code was “decommissioned” and maybe never archived or released to the public.)

various alignments and massive themes going on here. the Chinese have funneled billions of dollars into AI research and yet this match was not really publicized much or celebrated within china, in stark contrast to last years international Korean media extravaganza, verging on near-media-circus! the state censorship system apparently decided to cut off the live feed of the match ~30m into it after uncertain/ unknown reasons, and/ or it became clear that maybe their human player would not dominate. and then theres the trickiness of Google having withdrawn from china in the prior few years due to their hacking, and wanting back into the gargantuan market. but oops! chinas semi-state-sponsored companies compete directly with google services. as the teenagers say, AWKWARD! 😳

Continue reading