Computer Science

Marcus Hutter (born 1967) is a German computer scientist and professor at the Australian National University. Hutter was born and educated in Munich, where he studied physics and computer science at the Technical University of Munich. In 2000 he joined Jürgen Schmidhuber's group at the Swiss Artificial Intelligence lab IDSIA, where he developed the first mathematical theory of optimal Universal Artificial Intelligence, based on Kolmogorov complexity and Ray Solomonoff's theory of universal inductive inference. In 2006 he also accepted a professorship at the Australian National University...

vzn

there does seem some connection.

yeah we chatted about him quite awhile ago.

he funds the hutter prize etc.

Realz Slaw

yes

vzn

mentioned loebner in contrast at the time.

personally, do think of a lot of complexity theory in terms of compression.

for example P=?NP seems to relate to circuits, and how circuits can compress computation/fns.

ever mention this q before? was musing on it awhile back. compression vs complexity class separations

6

Q: compression of a Turing machine run sequence

consider a Turing machine with a set of states $s_n$ and alphabet symbols $a_n$. now consider a "run sequence" generated from a starting input in the following sense. the run sequence is defined as the sequence of state-symbol pairs that ensue in the computation. call the $i$th ensuing state $s'_...

not surprisingly the answer cites kolmogorov fn K(x)

btw am interested in funding open sci with bitcoin. whaddya think?

Realz Slaw

5:44 PM

heh

sure

who has the bitcoin :D

vzn

yeah exactly. (satoshi nakamoto?) :p

there are some kickstarter-like bitcoin analogs

Realz Slaw

sure

vzn

14

Q: Is there a crowdfunding site for Bitcoin?

Does a crowdfunding site like kickstarter.com exist, which accepts bitcoins instead of dollars?

websites finance

dude, have 1 q for you ... or maybe already asked this.

you use a pseudonym. so why the concern about googling stuff in chat rooms?

Realz Slaw

it is not an anonymous pseudonym

it can be easily connected to me heh

vzn

"it" seems so, it has very little info associated with it

Realz Slaw

5:47 PM

also I just don't want the content to be googlable

trust me, I can dox myself :D

vzn

ah yes the doxing thing

a new verb neologism invented by hackers

Realz Slaw

not hackers

most doxers are kiddies

anyone + google can be doxing

vzn

ok

funky

doxing, urban dictionary

mentions hackers specifically :p

Realz Slaw

/sigh

there is no hacking involved 99% of the time

EOS

vzn

EOS?

Realz Slaw

5:51 PM

end of story

vzn

ok

Realz Slaw

you can link me 1000 other sites, I wouldn't care

vzn

RJL has cranked out 2 blogs already in only a few days, another wonder

Realz Slaw

lol

vzn

told friend about you & your smarts, the friend wonders why you left college

Realz Slaw

5:54 PM

complicated

vzn

told friend, colleges are $$$ these days

heh yes

youre in US?

yes

NY?

5:55 PM

yeah

vzn

NYC cool. visited ~3 times.

Realz Slaw

heh ok

vzn

what do you think of satoshi nakamoto "outing"

Realz Slaw

I felt bad for the poor guy, then felt bad for the poor reporter

vzn

yeah

Realz Slaw

6:22 PM

@vzn pretty sure i finally figured out how to encode factoring into SAT with a linearithmic blowup

and practical

i.e better than karatsuba or somesuch

which has been my hobby for two years :D

and this is during research for something else :D

vzn

6:34 PM

sounds significant

Realz Slaw

maybe

I'll have to actually try it

vzn

think factoring->SAT not very well explored given its significance. felt that way 2decades ago also :\

Realz Slaw

well it is explored naturally

because every circuit is basically SAT

so if you try to figure out why the circuit is hard to reverse

vzn

think about it like this

Realz Slaw

you are essentially doing research on why the SAT is hard to reverse

vzn

6:37 PM

arithmetic/math approaches are a specialized/maybe narrow subset of all operations that analyze circuits.

that is also the idea behind "SAT induction"...

think big data/graph based approaches may yield new general analysis methods.

Realz Slaw

nah big graph is usually inexact stuff

vzn

found a large SAT masters thesis that tends to support this & go in that direction.

Realz Slaw

statistical

it doesn't apply IMO

vzn

SAT circuits are big graphs. etc.

Realz Slaw

they are

vzn

6:38 PM

it doesnt apply yet wink

Realz Slaw

but you are seeking exact answers

vzn

true.

Realz Slaw

the entire point of big data is to relax the exactness to get useful overall ideas about teh data

vzn

think some big data approaches will yield exact answers

true much big data approaches are statistical.

but the relaxation of the factoring->SAT graph may yield insight into exact structures.

& other graphs related to deep math problems etc.

RJL latest blog shows deep links between eg hamming coding theory & prime distributions etc

admit its still all very sketchy.

its a very longterm research program :\

decades, possibly.

Realz Slaw

mmm

vzn

6:42 PM

forgot to cite hilberts 10th problem on RJLs blog wrt that hamming coding post, oops.

hilberts 10th problem over century old & basically still being explored.

with new insight.

one of the greatest problems of the 20th century but still leaking into the 21st...

vzn

6:59 PM

re stdized testing... opting out

ORBOT Inc.

9:23 PM

Question: How can we show that there cannot be an heap with amoritzed O(1) insert/delete-min/get-min without using reduction to sorting?

Transcript for