Tag Archives: amusing

Items of some interest:

Posted on by

These are my recent Pin​board​.in links:

  • What if Inter­ac­tiv­ity is the New Pas­siv­ity? Jonathan Sterne /​ McGill Uni­ver­sity | Flow

    “What if all the bad things that media crit­ics have been said about pas­siv­ity for the past cen­tury or two are now equally applic­a­ble to all the demands to inter­act, to par­tic­i­pate? What if inter­ac­tiv­ity is now one of the cen­tral hinges through which power works? In many moments today, the most com­pli­ant ges­ture we can make is to con­sent to inter­act on the terms pre­sented to us by our soft­ware and machines. This pull is espe­cially strong in those com­mer­cial plat­forms that cel­e­brate their own dif­fer­ence from the so-​​called pas­sive media of pre­vi­ous decades, and in the process mon­e­tize their users’ par­tic­i­pa­tion either directly or indi­rectly. What if—from time to time—we chose not to iden­tify with the inter­ac­tive promise of new media plat­forms or for that mat­ter new media art? What if, when the new media savants lam­bast so-​​called old media audi­ences as denizens of pas­siv­ity and ide­ol­ogy, we say, “yes, that’s me”?”

    a-​​bit-​​too-​​theoryish cultural-​​norms ingroup-​​outgroup new-​​media

  • How Can Her­bert Spencer’s 1892 Revi­sions to his Social Sta­t­ics Help Us Under­stand Con­ser­v­a­tive Oppo­si­tion to the Indi­vid­ual Man­date? | Rortybomb

    “But I think it’s clear what his real objec­tion was: uni­ver­sal suf­frage has the poten­tial to advance social­is­tic causes, inter­fer­ing with his laissez-​​faire project. From his auto­bi­og­ra­phy: “Another exten­sion of the fran­chise since made…will inevitably be fol­lowed by a still more rapid growth of social­is­tic leg­is­la­tion.” When he real­ized women’s equal­ity could poten­tially inter­fere with laissez-​​faire eco­nom­ics, it was time for women’s equal­ity to get cut from his over­all the­ory of a bet­ter world. He would rather muti­late his intel­lec­tual project instead of allow­ing his ene­mies to con­tinue to build their gov­er­nance project.”

    Herbert-​​Spencer laissez-​​faire cor­po­ratism cap­i­tal­ism pol­i­tics con­ser­vatism via:cshalizi

  • BloJJ — About con­fer­ence poster design and defense:

    “My approach is dif­fer­ent. Poster pre­sen­ta­tion, like con­fer­ence pre­sen­ta­tion, belongs more to the area of dra­matic arts than to mar­ket­ing. It is information/​entertainment, and that is the main thing you have to bear in mind when prepar­ing for the ses­sion. Plus, while at a con­fer­ence you have the full atten­tion of your audi­ence (shared, of course, with email, Face­book, plus the 10% that are sim­ply speak­ing) in a poster ses­sion you have to first attract the atten­tion of the peo­ple wan­der­ing around a hall shared with other 20 to 100 posters, then keep them there for the dura­tion of the spiel and while you start a new one, and then, of course, con­vey the infor­ma­tion you want to share with your poster. ”

    advice academic-​​culture meet­ing poster-​​presentaitons skills

  • Economist’s View: The 999

    “Some Indi­vid­u­als of our Coun­try­men, by the Smiles of Prov­i­dence or some other Means, are enabled to roll in their four–wheel’d Car­riages, and can sup­port the Expence of good Houses, rich Fur­ni­ture, and Lux­u­ri­ous Liv­ing. But, is it equi­table that 99, or rather 999 should suf­fer for the Extrav­a­gance or Grandeur of one? Espe­cially when it is consider’d, that Men fre­quently owe their Wealth to the Impov­er­ish­ment of their Neighbours.”

    it-​​was-​​ever-​​thus

  • Ris­ingTide­Har­bor: Matt Barcomb’s Blog on Lean Agile Busi­ness Soft­ware Devel­op­ment: Stop B*tching About Local Optimizations

    “In fact, one approach is to inten­tion­ally over opti­mize a local opti­miza­tion. This will often make appar­ent to man­age­ment (or even to you) where the true bot­tle neck in the sys­tem is. We shouldn’t worry so much about doing the wrong things righter, but we should be aware that that may be the case and always work to be doing the right things. In the end, show­ing improve­ment and build­ing momen­tum can lead to excit­ing changes. In fair­ness, it can also come crash­ing to the ground if the right kinds of changes aren’t made at some point, but this should not deter any­one who thinks some­thing can be made bet­ter from try­ing to do so and it cer­tainly should not be a rea­son to do nothing!”

    change cultural-​​engineering organizational-​​behavior local-​​optimization

  • Geof­frey Chaucer Hath a Blog: A Long Tyme Agoon in a Shire Far Away

    “…A WHINY YOUTHE cam nexte, barl­eye a man, With yelwe haire, tunique, and farmeres tan. But aqua­cul­ture litel did he love, He wolde been a pilot al above And bulls­eye oump-​​rattes yn a nim­ble craft.…”

    amus­ing

  • knitr: Ele­gant, flex­i­ble and fast dynamic report gen­er­a­tion with R | knitr

    “The knitr pack­age was designed to be a trans­par­ent engine for dynamic report gen­er­a­tion with R, solve some long-​​standing prob­lems in Sweave, and com­bine fea­tures in other add-​​on pack­ages into one pack­age (knitr ≈ Sweave + cacheSweave + pgf­Sweave + weaver + R2HTML::RweaveHTML + highlight::HighlightWeaveLatex + 0.2 * brew + 0.1 * SweaveListingUtils + more).”

    R-​​language LaTeX type­set­ting dynamic-​​documents writ­ing tools

  • Math Notes | Futil­ity Closet

    kata for GP

    nudge-​​targets mathematical-​​recreations

  • Cere­bral Mastication

    “There’s a charm­ing lit­tle brain teaser that’s going around the Inter­webs. It’s got var­i­ous forms, but they all look some­thing like this:…”

    nudge-​​targets mathematical-​​recreations

  • Tanya Khovanova’s Math Blog » Blog Archive » Inter­lock­ing Polyominoes

    “A set of poly­omi­noes is inter­locked if no sub­set can be moved far away from the rest. It was known that poly­omi­noes that are built from four or fewer squares do not inter­lock. The project of Dhawan and his men­tor was to inves­ti­gate the inter­locked­ness of larger poly­omi­noes. And they totally deliv­ered. They quickly proved that you can inter­lock poly­omi­noes with eight or more squares. Then they proved that pen­tomi­noes can’t inter­lock. This left them with a gray area: what hap­pens with poly­omi­noes with six or seven squares? After draw­ing many beau­ti­ful pic­tures, they finally found the struc­ture pre­sented in our accom­pa­ny­ing image. The sys­tem con­sists of 12 hex­omi­noes and 5 pen­tomi­noes, and it is rigid. You can­not move a thing. That means that hex­omi­noes can be inter­locked and thus the gray area was resolved.”

    poly­omi­noes mathematical-​​recreations nudge-​​targets

  • Pool based evo­lu­tion­ary algo­rithm pre­sented in EvoStar 2012 « GeNeura Team

    “This is the first inter­na­tion­ally pub­lished paper (it was pre­vi­ously pub­lished in a Span­ish con­fer­ence of a series that deals with a sys­tem, intended for vol­un­teer com­put­ing, that uses a pool for imple­ment­ing dis­trib­uted evo­lu­tion­ary algo­rithms. The basic idea is that the pop­u­la­tion resides in a pool (imple­mented using CouchDB), with clients pulling indi­vid­u­als from the pool, doing stuff on them, and putting them back in the pool. The algo­rithm uses, as much as pos­si­ble, CouchDB fea­tures (such as revi­sions and views) to achieve good per­for­mance. All the code (for this and, right now, for the next papers) is avail­able as open-​​source code.”

    distributed-​​processing evolutionary-​​algorithms CouchDB nudge

  • What Amazon’s ebook strat­egy means — Charlie’s Diary

    “If the major pub­lish­ers switch to sell­ing ebooks with­out DRM, then they can enable cus­tomers to buy books from a vari­ety of out­lets and move away from the walled gar­den of the Kin­dle store. They see DRM as a defense against piracy, but piracy is a much less imme­di­ate threat than a gigan­tic multi­na­tional with rev­enue of $48 Bil­lion in 2011 (more than the entire global pub­lish­ing indus­try) that has expressed its inten­tion to “dis­rupt” them, and whose chief exec­u­tive said recently “even well-​​meaning gate­keep­ers slow inno­va­tion” (where “inno­va­tion” is code-​​speak for “oppor­tu­ni­ties for me to turn a profit”). And so they will deep-​​six their exist­ing com­mit­ment to DRM and use the terms of the DoJ-​​imposed set­tle­ment to wig­gle out of the most-​​favoured-​​nation terms imposed by Ama­zon, in order to sell their wares as widely as pos­si­ble. If they don’t, they’re doomed. And all of us who like to read (or write) fic­tion get to live in the Ama­zon com­pany town.”

    monopoly-​​and-​​monpsony-​​sittin-​​in-​​a-​​tree Ama­zon eBooks disintermediation-​​in-​​action cor­po­ratism redis­in­ter­me­di­a­tion

Items of some interest:

Posted on by

These are my recent Pin​board​.in links:

Items of some interest…

Posted on by

These are my recent Pin​board​.in links:

  • [1109.5664] Deter­min­is­tic Fea­ture Selec­tion for $k$-means Clustering

    “We study fea­ture selec­tion for $k$-means clus­ter­ing. Although the lit­er­a­ture con­tains many meth­ods with good empir­i­cal per­for­mance, algo­rithms with prov­able the­o­ret­i­cal behav­ior have only recently been devel­oped. Unfor­tu­nately, these algo­rithms are ran­dom­ized and fail with, say, a con­stant prob­a­bil­ity. We address this issue by pre­sent­ing a emph{deterministic} fea­ture selec­tion algo­rithm for $k$-means with the­o­ret­i­cal guar­an­tees. At the heart of our algo­rithm lies a deter­min­is­tic method for decom­po­si­tions of the identity.”

    clus­ter­ing sta­tis­tics algo­rithms nudge-​​targets

  • [1110.5190] Constant-​​factor approx­i­ma­tion of dom­i­na­tion num­ber in sparse graphs

    “The k-​​domination num­ber of a graph is the min­i­mum size of a set X such that every ver­tex of G is in dis­tance at most k from X. We give a lin­ear time constant-​​factor approx­i­ma­tion algo­rithm for k-​​domination num­ber in classes of graphs with bounded expan­sion, which include e.g. proper minor-​​closed graph classes, classes closed on topo­log­i­cal minors or classes of graphs that can be drawn on a fixed sur­face with bounded num­ber of cross­ings on each edge. The algo­rithm is based on the fol­low­ing approx­i­mate min-​​max char­ac­ter­i­za­tion. A sub­set A of ver­tices of a graph G is d-​​independent if the dis­tance between each pair of ver­tices in A is greater than d. Note that the size of the largest 2k-​​independent set is a lower bound for the k-​​domination num­ber. We show that every graph from a fixed class with bounded expan­sion con­tains a 2k-​​independent set A and a k-​​dominating set D such that |D|=O(|A|), and these sets can be found in lin­ear time. For dom­i­na­tion num­ber (k=1) the assump­tions can be relaxed, and the result holds for all graph classes with arrange­abil­ity bounded by a constant.”

    operations-​​research com­bi­na­torics graph-​​theory algo­rithms nudge-​​targets

  • [1112.1945] Approx­i­ma­tion Algo­rithms for Edge Par­ti­tioned Ver­tex Cover Problems

    “In the Par­tial Ver­tex Cover (PVC) prob­lem we are given an undi­rected graph G = (V, E), a pos­i­tive cost asso­ci­ated with each ver­tex and a pos­i­tive inte­ger k and the goal is to find a min­i­mum cost sub­set of ver­tices S such that atleast k edges of the graph are cov­ered. In this paper we con­sider two new gen­er­al­iza­tion of the PVC prob­lem. In the first vari­a­tion which we call Par­ti­tion Ver­tex Cover (Partition-​​VC) prob­lem, the edges of the graph G are divided into n dis­joint par­ti­tions $P_​1, P_​2… P_​n$ and we have to select a min­i­mum cost sub­set of ver­tices S such that atleast $k_​i$ edges are cov­ered from par­ti­tion $P_​i$. In the sec­ond vari­a­tion which we call Knap­sack Par­ti­tion Ver­tex Cover (KPVC) prob­lem, in addi­tion to the pre­vi­ous con­di­tions, each edge e has a profit $pi_{e}$ asso­ci­ated with it and we have an added knap­sack con­straint that the total profit of the cov­ered edges in par­ti­tion $P_​i$ should be atleast $Pi_​i$. We give an $O(log n)$ approx­i­ma­tion for both the prob­lems using a com­bi­na­tion of deter­min­is­tic round­ing and ran­dom­ized round­ing approach that oper­ates on the LP strength­ened by adding Knap­sack Cover inequal­i­ties as pro­posed by Carr, Fleis­cher, Leung & Phillips. We also show that these bounds can not be fur­ther improved by reduc­ing the set cover prob­lem to the Partition-​​VC prob­lem in poly­no­mial time. We also give an $O(f)$ approx­i­ma­tion for the Partition-​​VC prob­lem using a pri­mal dual schema where f is the max­i­mum num­ber of edges in any partition.”

    operations-​​research graph-​​theory graph-​​partitioning linear-​​programming nudge-​​targets

  • [1101.3501] Con­ver­gence rates of effi­cient global opti­miza­tion algorithms

    “Effi­cient global opti­miza­tion is the prob­lem of min­i­miz­ing an unknown func­tion f, using as few eval­u­a­tions f(x) as pos­si­ble. It can be con­sid­ered as a continuum-​​armed ban­dit prob­lem, with noise­less data and sim­ple regret. Expected improve­ment is per­haps the most pop­u­lar method for solv­ing this prob­lem; the algo­rithm per­forms well in exper­i­ments, but lit­tle is known about its the­o­ret­i­cal prop­er­ties. Imple­ment­ing expected improve­ment requires a choice of Gauss­ian process prior, which deter­mines an asso­ci­ated space of func­tions, its reproducing-​​kernel Hilbert space (RKHS). When the prior is fixed, expected improve­ment is known to con­verge on the min­i­mum of any func­tion in the RKHS. We begin by pro­vid­ing con­ver­gence rates for this pro­ce­dure. The rates are opti­mal for func­tions of low smooth­ness, and we mod­ify the algo­rithm to attain opti­mal rates for smoother func­tions. For prac­ti­tion­ers, how­ever, these results are some­what mis­lead­ing. Pri­ors are typ­i­cally not held fixed, but depend on para­me­ters esti­mated from the data. For stan­dard esti­ma­tors, we show this pro­ce­dure may never dis­cover the min­i­mum of f. We then pro­pose alter­na­tive esti­ma­tors, cho­sen to min­i­mize the con­stants in the rate of con­ver­gence, and show these esti­ma­tors retain the con­ver­gence rates of a fixed prior.”

    opti­miza­tion operations-​​research theory-​​and-​​practice-​​sitting-​​in-​​a-​​tree nudge-​​targets algo­rithms

  • [1011.1939] Dis­crete Par­ti­tion­ing and Cov­er­age Con­trol for Gos­sip­ing Robots

    “We pro­pose dis­trib­uted algo­rithms to auto­mat­i­cally deploy a team of mobile robots to par­ti­tion and pro­vide cov­er­age of a non-​​convex envi­ron­ment. To han­dle arbi­trary non-​​convex envi­ron­ments, we rep­re­sent them as graphs. Our par­ti­tion­ing and cov­er­age algo­rithm requires only short-​​range, unre­li­able pair­wise “gos­sip” com­mu­ni­ca­tion. The algo­rithm has two com­po­nents: (1) a motion pro­to­col to ensure that neigh­bor­ing robots com­mu­ni­cate at least spo­rad­i­cally, and (2) a pair­wise par­ti­tion­ing rule to update ter­ri­tory own­er­ship when two robots com­mu­ni­cate. By study­ing an appro­pri­ate dynam­i­cal sys­tem on the space of par­ti­tions of the graph ver­tices, we prove that ter­ri­tory own­er­ship con­verges to a pairwise-​​optimal par­ti­tion in finite time. This new equi­lib­rium set rep­re­sents improved per­for­mance over com­mon Lloyd-​​type algo­rithms. Addi­tion­ally, we detail how our algo­rithm scales well for large teams in large envi­ron­ments and how the com­pu­ta­tion can run in any­time with lim­ited resources. Finally, we report on large-​​scale sim­u­la­tions in com­plex envi­ron­ments and hard­ware exper­i­ments using the Player/​Stage robot con­trol system.”

    com­plex­ol­ogy robot­ics agent-​​based computational-​​geometry nudge-​​targets voronoi emergent-​​design

  • [1112.1841] Con­sis­tency of mul­ti­di­men­sional com­bi­na­to­r­ial substitutions

    “Mul­ti­di­men­sional com­bi­na­to­r­ial sub­sti­tu­tions are rules that replace sym­bols by finite pat­terns of sym­bols in Z^d. We focus on the case where the pat­terns are not nec­es­sar­ily rec­tan­gu­lar, which requires a spe­cific descrip­tion of the way they are glued together in the image by a sub­sti­tu­tion. Two prob­lems can arise when defin­ing a sub­sti­tu­tion in such a way: it can fail to be con­sis­tent, and the pat­terns in an image by the sub­sti­tu­tion might over­lap. We prove that it is unde­cid­able whether a two-​​dimensional sub­sti­tu­tion is con­sis­tent or over­lap­ping, and we pro­vide prac­ti­cal algo­rithms to decide these prop­er­ties in some par­tic­u­lar cases.”

    frac­tals rewriting-​​systems mathematical-​​recreations amus­ing nudge-​​targets unde­cod­abil­ity

  • [1112.3523] Approx­i­mat­ing the Edge Length of 2-​​Edge Con­nected Pla­nar Geo­met­ric Graphs on a Set of Points

    “Given a set $P$ of $n$ points in the plane, we solve the prob­lems of con­struct­ing a geo­met­ric pla­nar graph span­ning $P$ 1) of min­i­mum degree 2, and 2) which is 2-​​edge con­nected, respec­tively, and has max edge length bounded by a fac­tor of 2 times the opti­mal; we also show that the fac­tor 2 is best pos­si­ble given appro­pri­ate con­nec­tiv­ity con­di­tions on the set $P$, respec­tively. First, we con­struct in $O(nlog{n})$ time a geo­met­ric pla­nar graph of min­i­mum degree 2 and max edge length bounded by 2 times the opti­mal. This is then used to con­struct in $O(nlog n)$ time a 2-​​edge con­nected geo­met­ric pla­nar graph span­ning $P$ with max edge length bounded by $sqrt{5}$ times the opti­mal, assum­ing that the set $P$ forms a con­nected Unit Disk Graph. Sec­ond, we prove that 2 times the opti­mal is always suf­fi­cient if the set of points forms a 2 edge con­nected Unit Disk Graph and give an algo­rithm that runs in $O(n^2)$ time. We also show that for $k in O(sqrt{n})$, there exists a set $P$ of $n$ points in the plane such that even though the Unit Disk Graph span­ning $P$ is $k$-vertex con­nected, there is no 2-​​edge con­nected geo­met­ric pla­nar graph span­ning $P$ even if the length of its edges is allowed to be up to 17⁄16.”

    graph-​​theory geom­e­try algo­rithms computational-​​geometry nudge-​​targets

Items of some interest…

Posted on by

These are my recent Pin​board​.in links:

  • Nelson’s Weblog: tech /​ foss4g-​​2011-​​trip-​​report

    “…The rest of this post is a link dump of some of the peo­ple and things I saw at the con­fer­ence. I’m no “cura­tor,” just a typ­ist, sorry for the lack of organization.”

    geo software-​​development map­ping resources

  • main — katt83project

    col­lec­tion of scripts and other tools in sup­port of the Dis­trib­uted Proof­read­ers workflow

    Distributed-​​Proofreaders script­ing toolkit

  • Presto Chango | Futil­ity Closet

    “It will be observed that this square when turned upside down is still magic.”

    mathematical-​​recreations amus­ing nudge-​​targets

  • Ruth Kinna on Guy Aldred | berfrois

    “Guy Aldred is an obscure but impor­tant fig­ure in the his­tory of social­ist thought. He some­times crops up in his­to­ries of British social­ism, syn­di­cal­ist and labour organ­i­sa­tion, but rarely in dis­cus­sions of social­ist the­ory. His uncom­pro­mis­ing com­mit­ment to activism per­haps explains this neglect: as Aldred him­self argued in a com­men­tary on British anar­chism, ide­olo­gies are too often shaped by the philo­soph­i­cal reflec­tions of edu­cated elites, leav­ing the thoughts of work­ing class auto­di­dacts who spend a life­time stand­ing on street cor­ners, pro­pa­gan­dis­ing, ignored. Per­haps, too, his evan­gel­i­cal roots make his work an acquired taste: Aldred writes with moral cer­tainty and con­vic­tion that leaves lit­tle room for debate. Most bio­graph­i­cal accounts sug­gest that he was not an easy man to get along with and though he did not lack organ­i­sa­tional skill, he found co-​​operation dif­fi­cult. The plea­sure he took in the pun of his name – ‘the man they all dread’ – was indica­tive of the prob­lem. Yet Aldred’s ideas are com­pelling and the judge­ments he made in his early life were con­sis­tently rev­o­lu­tion­ary, lib­er­tar­ian, anar­chis­tic and usu­ally good. Aldred cam­paigned against mar­riage and for birth con­trol in sup­port of women’s lib­er­a­tion before the First World War; he encour­aged con­sci­en­tious objec­tion in both world con­flicts and pub­li­cised the vin­dic­tive abuse that COs suf­fered for tak­ing their stance. In all his early writ­ings, he ele­vated the strug­gles of com­mon peo­ple – from reli­gious non-​​conformists to con­victs. Draw­ing on the reports of his com­rades, Ethel Mac­Don­ald (1909−1960) and Jane (Jenny) Patrick (1884−1971), he sup­ported the 1936 anar­chist rev­o­lu­tion in Spain [1] and until his later life, he con­sis­tently opposed the dog­ma­tism of ortho­dox Marx­ism, whether it was expressed in the the­o­ret­i­cal pieties of the Euro­pean social demo­c­ra­tic move­ment or, after the Russ­ian Rev­o­lu­tion, in the cold, phys­i­cal bru­tal­ity of the Stal­in­ist regime. The pas­sion with which he advanced these causes cap­tures the spirit of an opti­mistic, utopian, roman­tic cur­rent of social­ism whose hopes and ideals, squeezed by social democ­racy on one side and state social­ism on the other, were ulti­mately dis­ap­pointed but which remain inspiring.”

    anar­chism his­tory biog­ra­phy social­ism

Items of some interest…

Posted on by

These are my recent Pin​board​.in links: