January 2019 - nexa - server-nexa.polito.it

Overlooked No More: Karen Sparck Jones, Who Established the Basis for Search Engines
by Alberto Cammozzo Jan. 15, 2019

Jan. 15, 2019

La donna che inventò l'inverse document frequency (TF/IDF è una misura fondamentale per chi fa analisi testuale). Sparck Jones [] came up with a slogan: “Computing is too important to be left to men.” [] Sparck Jones cautioned engineers to think of their work’s impact on society. “There is an interaction between the context and the programming task itself,” she said. “You don’t need a fundamental philosophical discussion every time you put finger to keyboard, but as computing is spreading so far into people’s lives, you need to think about these things.” <https://www.nytimes.com/2019/01/02/obituaries/karen-sparck-jones-overlooked…> Overlooked No More: Karen Sparck Jones, Who Established the Basis for Search Engines A pioneer of computer science for work combining statistics and linguistics, and an advocate for women in the field. <https://www.nytimes.com/2019/01/02/obituaries/karen-sparck-jones-overlooked…> /Since 1851, obituaries in The New York Times have been dominated by white men. With //Overlooked/ <https://nytimes.com/overlooked>/, we’re adding the stories of remarkable people whose deaths went unreported in The Times./ By Nellie Bowles When most scientists were trying to make people use code to talk to computers, Karen Sparck Jones taught computers to understand human language instead. In so doing, her technology established the basis of search engines like Google. A self-taught programmer with a focus on natural language processing, and an advocate for women in the field, Sparck Jones also foreshadowed by decades Silicon Valley’s current reckoning, warning about the risks of technology being led by computer scientists who were not attuned to its social implications. “A lot of the stuff she was working on until five or 10 years ago seemed like mad nonsense, and now we take it for granted,” said John Tait, a longtime friend who works with the British Computer Society. Sparck Jones’s seminal 1972 paper in the Journal of Documentation laid the groundwork for the modern search engine. In it, she combined statistics with linguistics — an unusual approach at the time — to establish formulas that embodied principles for how computers could interpret relationships between words. By 2007, Sparck Jones said, “pretty much every web engine uses those principles.” “Anything that does index-term weighting using any kind of statistical information will be using a weighting function that I published in 1972,” she said in an interview with the British Computer Society <https://www.bcs.org/content/ConWebDoc/10791>. Karen Ida Boalth Sparck Jones was born on Aug. 26, 1935, in Huddersfield, England, a textile manufacturing town. Her parents were Alfred Owen Jones, a chemistry lecturer, and Ida Sparck, who worked for the Norwegian government while in exile in London during World War II. When studying history and then philosophy (the department was then called moral sciences) at Cambridge, she met the head of the Cambridge Language Research Unit, Margaret Masterman, who would inspire her to enter the field. Sparck Jones later describe <http://ethw.org/Oral-History:Karen_Sp%C3%A4rck_Jones>d her as “a very strange and interesting woman” who, unusual for the time, used her maiden name professionally. Sparck Jones, too, kept her name when she married Roger Needham <https://www.nytimes.com/2003/03/06/business/roger-needham-computer-security…>, a fellow computer scientist, in 1958, saying, “It maintains a permanent existence of your own.” Sparck Jones started working for Ms. Masterman. She wanted to figure out how to program a computer to understand words that could have many meanings (for example “field”) and set about programming a massive thesaurus. “A lot of the stuff she was working on until five or 10 years ago seemed like mad nonsense, and now we take it for granted,” John Tait, of the British Computer Society, said about Sparck Jones.CreditComputer Laboratory/University of Cambridge“All words in a natural language are ambiguous; they have multiple senses,” she said in anoral history interview for the History Center of the Institute of Electrical and Electronics Engineers. “How do you find out which sense they’ve got in any particular use?” In 1964, Sparck Jones published “Synonymy and Semantic Classification,” which is now seen as a foundational paper in the field of natural language processing. In 1972, she introduced the concept of inverse document frequency, which counts the number of times a term is used in a document in order to determine the term’s importance; it, too, is a foundation of modern search engines. Sparck Jones began working on early speech recognition systems in the 1980s. Most mornings and afternoons, she and her husband, a pioneer in software security, debated theory in the Cambridge department’s tea room. Their home in Coton, just west of Cambridge, was full of books, art and found items, like an interesting piece of driftwood or a Victorian-era knife grinder. They had a second house in the same village, using it to store the overflow of their book collection and as her artist’s arch Lab. Sailing was another passion of Sparck Jones and Needham. They restored an 1872 vintage sailboat called Fanny of Cowes <http://www.oga.org.uk/boat/fanny-cowes> and raced it against other old boats along the east coast of England. They chose not to have children. “They wanted their intellectual life,” said Andrew Herbert, her friend and a fellow computer scientist. “They were clearly deeply in love with each other all the way through their life.” Sparck Jones had a booming voice and a puckish sense of humor. At work, she usually wore a simple uniform: bluejeans, red sweater, white blouse. She also wore a brooch, which she made from stones and part of a horseshoe. When she had to bike to a formal dinner, as one often did at Cambridge, she was known to use clothing pegs to pin her dress to the handlebars. In 1982, the British government tapped Sparck Jones to work on the Alvey Program, an initiative to encourage more computer science research across the country. In 1993, she wrote, with Julia R. Galliers, “Evaluating Natural Language Processing Systems,” the seminal textbook on the topic. <https://www.springer.com/us/book/9783540613091> Sparck Jones became president of the Association for Computational Linguistics, an international group for professionals in the field, in 1994. She became a full-time professor at Cambridge in 1999 — and it had bothered her that it took so long. For all the years before, she had been on contract with the university, an untenured and lower-status form of academic employment referred to as “living on soft money.” “Cambridge was in many ways not user-friendly, in the sense of women-friendly,” she said of the delay. Sparck Jones died of cancer on April 4, 2007. She was 71. Though she did not receive an obituary in The Times, her husband did, in 2003 <https://www.nytimes.com/2003/03/06/business/roger-needham-computer-security…>. Today, researchers are still citing her formulas. Ideas she wrote about are now being put into practice as artificial intelligence research becomes more prevalent. “It points to how far ahead of her time she was, how consequential her work was, how little it was valued for the first 20 years,” said Martha Palmer, a professor in the Linguistics and Computer Science departments at the University of Colorado. Sparck Jones mentored a generation of researchers, male and female,and came up with a slogan: “Computing is too important to be left to men.” She was ahead of her time in another respect. Decades before Silicon Valley was having its moral reckoning, Sparck Jones cautioned engineers to think of their work’s impact on society. “There is an interaction between the context and the programming task itself,” she said. “You don’t need a fundamental philosophical discussion every time you put finger to keyboard, but as computing is spreading so far into people’s lives, you need to think about these things.”

1 0

Article 13 is almost finished – and it will change the internet as we know it
by Giacomo Tesio Jan. 15, 2019

Jan. 15, 2019

Da https://juliareda.eu/2019/01/article-13-almost-finished/ ``` Negotiations about the EU copyright reform law have resumed: After missing the original Christmas deadline, negotiators for the European Parliament and Council are now aiming to finalise the text on January 21, 2019. The negotiators have reached agreement on the core of Article 13, which will change the internet as we know it: They want to make internet platforms directly liable for any copyright infringements their users commit. What remains to be decided: Exactly what lengths will platforms need to go to to avoid or limit their liability? Just how much they will need to restrict our ability to post and share our creations online? ``` Oltre 10 anni fa, se non ricordo male, fu trafugato parte del codice sorgente di Windows (credo che fosse XP, ma vado a memoria). Sarebbe divertente vedere Microsoft responsabile della violazione del proprio copyright da parte di un utente se quel sorgente fosse uploadato su GitHub... :-) Giacomo

1 0

Il matching algoritmico al potere: Gilets Jaunes e populismo social
by Antonio Casilli Jan. 14, 2019

Jan. 14, 2019

La Fondation Jean-Jaurès (vicina a quel che resta del Parti Socialiste, quindi sappiate applicare le ponderazioni politiche del caso) pubblica nell'Obs (settimanale centrista) una lunga nota in due parti sulla genealogia digitale del movimento dei Gilets Jaunes. Une prima parte presenta una carrellata storica delle pagine Facebook più frequentate e si sofferma sul ruolo dell'algoritmo di Facebook nell'impennata della visibilità del movimento: > "Questo riorientamento emotivo e geografico nella prioritizzazione delle informazioni su Facebook è un importante punto di svolta. Il fatto che Priscillia > Ludosky ed Eric Drouet, i due utilizzatori dietro l'iniziativa di mobilitazione online, provengano dallo stesso dipartimento non è una coincidenza. Non si > conoscevano: è stato il nuovo algoritmo di Facebook che li ha messi in contatto, individuando interessi comuni e vicinanza spaziale. La diffusione del > movimento nelle province ha seguito la stessa logica e lo stesso meccanismo." https://www.nouvelobs.com/politique/20190111.OBS8351/plongee-dans-les-compt… La seconda parte contiene invece un'analisi dettagliata delle strategie di comunicazione sui social di Eric Drouet e Maxime Nicolle, due leader di fatto, parecchio orientati a destra, di un movimento che si vuole "senza leader" e "né di destra né di sinistra". > "Anche se preferiscono definirsi come semplici "messaggeri delle rivendicazioni popolari", presentano in realtà tutti gli attributi dei leader 2.0: liberi > da intermediari, non eletti, mai ufficialmente indotti, ma legittimati dal pubblico generato dai loro video. A questo proposito, i loro ripetuti appelli a > non nominare dei dirigenti suonano sempre più come il desiderio di impedire l'emergere di un nome che non sarebbe il loro." https://www.nouvelobs.com/politique/20190114.OBS8414/eric-drouet-et-maxime-…

1 0

"Automating Society"
by fabio chiusi Jan. 14, 2019

Jan. 14, 2019

Segnalo perché credo in lista possa interessare: il 29 gennaio viene presentato al Parlamento Europeo un rapporto di Algorithm Watch sull'automated decision-making (ADM) in Europa, e in 12 paesi in particolare (io ho curato la parte sull'Italia). Sul rapporto: "For the report “Automating Society – Taking Stock of Automated Decision-Making in the EU”, our experts have looked at the situation at the EU level but also in 12 Member States: Belgium, Denmark, Finland, France, Germany, Italy, Netherlands Poland, Slovenia, Spain, Sweden and the UK. We assessed not only the political discussions and initiatives in these countries but also present a section “ADM in Action” for all states, listing examples of automated decision-making already in use." Altre informazioni, qui: https://algorithmwatch.org/en/automating-society/ f.

1 0

Internet Governance Forum, i cittadini possono disegnare il futuro della rete
by Arturo Di Corinto Jan. 14, 2019

Jan. 14, 2019

Per i nexiani, da Repubblica-- Internet Governance Forum, i cittadini possono disegnare il futuro della rete È online la call per partecipare alla definizione dei temi e delle proposte da portare a Berlino all'Internet Governance Forum, il "parlamento" che ogni anno discute a livello mondiale la direzione dello sviluppo della rete. E l'Italia prova a organizzarsi *di ARTURO DI CORINTO* https://www.repubblica.it/tecnologia/sicurezza/2019/01/09/news/internet_gov… Aggiunta importante. Poche ore fa è venuta a mancare una nostra amica, Antonella Giulia Pizzaleo, con cui abbiamo condiviso un pezzo di percorso dell'IGF. L'articolo, scritto prima della notizia, lo dedico a lei. A. -- Arturo Di Corinto www.dicorinto.it ---------------------------------> AVVERTENZE AI SENSI DEL D.Lgs. 196/2003 Le informazioni contenute in questo messaggio di posta elettronica e/o nel/i file/s allegato/i sono da considerarsi strettamente riservate. Il loro utilizzo è consentito esclusivamente al destinatario del messaggio, per le finalità indicate nel messaggio stesso. Qualora riceviate questo messaggio senza esserne il destinatario, Vi preghiamo cortesemente di darcene notizia via e-mail e di procedere alla distruzione del messaggio stesso, cancellandolo dal Vostro sistema. Conservare il messaggio stesso, divulgarlo anche in parte, distribuirlo ad altri soggetti, copiarlo, od utilizzarlo per finalità diverse, costituisce comportamento contrario ai principi dettati dal D.Lgs. 196/2003

2 2

Don’t believe the hype: the media are unwittingly selling us an AI fantasy | John Naughton | Opinion | The Guardian
by J.C. DE MARTIN Jan. 13, 2019

Jan. 13, 2019

https://www.theguardian.com/commentisfree/2019/jan/13/dont-believe-the-hype… (Sent from my wireless device; please excuse brevity and typos (if any))

3 3

MISE: Selezionati gli esperti dei gruppi di Intelligenza Artificiale e Blockchain
by Antonio Vetro' Jan. 12, 2019

Jan. 12, 2019

Notizia pubblicata sul sito del MISE già una settimana fa, mi scuso qualora fosse già passata in lista: “ Intelligenza artificiale e blockchain: selezionati gli esperti Elaboreranno la strategia nazionale sull’intelligenza artificiale e la strategia nazionale in materia di tecnologie basate su registri condivisi e blockchain” - Giovedì, 27 Dicembre 2018 http://adisurv.sviluppoeconomico.gov.it/index.php/it/per-i-media/notizie/it… Qui i link per accedere agli elenchi degli esperti Intelligenza Artificiale https://www.mise.gov.it/index.php/it/10-istituzionale/ministero/2038906-int… <https://www.mise.gov.it/index.php/it/10-istituzionale/ministero/2038906-int…> Blockchain https://www.mise.gov.it/index.php/it/10-istituzionale/ministero/2039024-blo… <https://www.mise.gov.it/index.php/it/10-istituzionale/ministero/2039024-blo…> Un caro saluto av

4 3

I was wrong about Google and Facebook: there’s nothing wrong with them (so say we all)
by Giacomo Tesio Jan. 11, 2019

Jan. 11, 2019

Dopo una breve carrellata di insospettabili "collaborazionisti", una conclusione orwelliana: > War is Peace, Freedom is Slavery, Surveillance is Privacy > > So, in light of the overwhelming support for surveillance capitalism by generally > well-respected organisations that say they work to protect our human rights, > privacy, and democracy, I have decided that I must be the one who’s wrong. > > If Google, Facebook, etc., were even half as bad as I make them out to be, these > organisations would not be partnering with or endorsing them. > > I just want to apologise for being such a “Negative Nancy” and take this > opportunity to thank the FSF, The Free Software Conservancy, Mozilla, Apple, > The Nordic Privacy Arena, AccessNow and RightsCon, Amsterdam Privacy > Week, CPDP, and GNOME for showing me the error of my ways. I now know, > thanks to their moral leadership on this issue, that it’s perfectly fine for people > working to protect human rights and democracy to take millions and billions from > companies like Google and Facebook and to partner with them. > > I must admit, I feel a little silly. The last five years would have been so much > easier if only I’d understood this earlier. > > Here’s wishing you all a world full of freedom sponsored by Google, > privacy sponsored by Facebook, and democracy sponsored by Palantir. https://ar.al/2019/01/11/i-was-wrong-about-google-and-facebook-theres-nothi… Nel caso della CopyleftConf, credo vi siano ragioni molto più pratiche per non accettare "l'aiuto" di chi è strutturalmente incompatibile con il copyleft, ma nonostante il sarcasmo moralista, i fatti che riporta Aral sono veri. Quando dico che l'open source è uno strumento di marketing, non intendo sminuirne il valore economico ma riportare semplicemente la sua funzione principale: servire esigenze commerciali di chi ne paga lo sviluppo. D'altro canto il software libero non è, in sé, garanzia di privacy o di sicurezza, e non c'è ragione di credere che tutti gli hacker che lo sviluppano siano buoni e interessati alla privacy dei loro utenti. Se è vero che Aral è piuttosto monotematico, ha ragione su un punto: un hacker non può far finta di non vedere per qualche soldo in più da spendere in tortine al rinfresco. Quindi anche se Google non fosse lì per calmierare la conferenza sul Copyleft perché non si spinga su terreni scomodi per i suoi prodotti, sarebbe forse appropriato non accettarne i soldi. Giacomo

1 0

Do social media bots have a right to free speech?
by Alberto Cammozzo Jan. 11, 2019

Jan. 11, 2019

<https://thebulletin.org/2019/01/do-social-media-bots-have-a-right-to-free-s…> People who tweet in their jobs—let’s say 21st century journalists, just for example—might say that writing two million tweets represents a daunting challenge. That’s the rough number, Twitter says, that a Russian-linked set of accounts cranked out on the 2016 US presidential race in just the 10 weeks leading up to the election. Of course, in that case, the prolific “authors” were a collection of about 50,000 automated accounts often called bots. A new law in California will soon force bots that engage in electioneering or marketing to declare their non-human identity. While the Kremlin agents who interfered in the US election likely wouldn’t be beholden to a state-level law in the United States, or deterred by it, domestic political campaigns and businesses might. For at least one constitutional scholar, that possibility raises this question: Do bots, like citizens, have that most sacred right enshrined in the First Amendment to the US Constitution, the right to free speech? Laurent Sacharoff, a law professor at the University of Arkansas, thinks the people programming bots may want US courts to answer that in the affirmative. Take a hypothetical bot that engages a voter around a shared concern like motherhood, for instance. “If it has to say, ‘Well look, I’m not really a mother, I’m a chatbot mother, a mother of other chatbots. And when I say I feel your pain, I don’t actually have feelings.’ That’s just not going to be very effective,” Sacharoff says. A court ruling on whether bots have First Amendment free speech rights remains in the realm of conjecture. Meanwhile, bots on Twitter are very real. The Pew Research Center conducted a study in 2017 that looked at 1.2 million tweets that contained links. The researchers found that 66 percent of links were posted by “suspected bots,” with the percentage growing much higher depending on the type of content that was being linked. Bots posted about 90 percent of links to news aggregators, for instance. When a gunman killed 11 people in a synagogue in the US city of Pittsburgh last fall, Robhat Labs, which analyzes bots during big news events, found that bots drove 23 percent of the twitter activity around the issue during one 24-hour period. Bots often amplify extremist views. Twitter claims it is making progress in combating “spammy or automated” accounts and is making it harder to spread fake election content. As court rulings continue to define the scope of the First Amendment, it has grown to include corporations. A 2010 Supreme Court decision in the notorious Citizens United case opened the doors for corporations and unions to spend unlimited money directly on electioneering. In the eyes of the court, the elimination of limits on campaign spending was meant to ensure free speech. Sacharoff says a central point underpinning the ruling was that voters could benefit from hearing the “message of a corporation.” The benefit, he says, is for the audience, but the right lies with the corporation. The same principal could hold for bots. “Even though bots are abstract entities, we might think of them as having free speech rights to the extent that they are promoting or promulgating useful information for the rest of us,” Sacharoff says. “That’s one theory of why a bot would have a First Amendment free speech right, almost independent of its creators.” Alternatively, the bots could just be viewed as direct extensions of their human creators. In either case—whether because of an independent right to free speech or because of a human creator’s right—Sacharoff says, “you can get to one or another nature of bots having some kind of free speech right.” In 1943, the Supreme Court ruled that public schools in West Virginia could not force students to recite the Pledge of Allegiance, finding that the requirement violated the First Amendment. After all, “to sustain the compulsory flag salute, we are required to say that a Bill of Rights which guards the individual’s right to speak his own mind left it open to public authorities to compel him to utter what is not in his mind,” Justice Robert H. Jackson wrote in the majority opinion. The so-called compelled speech doctrine was born. “These are cases that say that the First Amendment prohibits the government not just from suppressing speech but also from compelling a person to state a message,” Sacharoff says. “And in this case, the message is, ‘I’m a bot, not a human being.’ If the bot or its human being behind it has a free speech right, then they have a right against being compelled to disclose that they’re bot.” In previous Bulletin coverage, the author of the new California law dismisses the idea that the law violates free speech rights. State Sen. Robert Hertzberg says anonymous marketing and electioneering bots are committing fraud. “My point is, you can say whatever the heck you want,” Hertzberg says. “I don’t want to control one bit of the content of what’s being said. Zero, zero, zero, zero, zero, zero. All I want is for the person who has to hear the content to know it comes from a computer. To me, that’s a fraud element versus a free speech element.” Sacharoff believes that the issue of bots and their potential First Amendment rights may one day have its day in court. Campaigns, he says, will find that bots are helpful and that their “usefulness derives from the fact that they don’t have to disclose that they’re bots.” “If some account is retweeting something, if they have to say, ‘I’m a bot’ every time, then it’s less effective. So sure I can see some campaign seeking a declaratory judgment that the law is invalid,” he says. “Ditto, I guess, [for] selling stuff on the commercial side.” In First Amendment cases, a court would consider whether there is a compelling reason to limit the scope of a person’s speech. In the case of electioneering social-media bots, the sanctity of elections would probably suffice, Sacharoff says. But then a court would weigh whether the anti-bot law is too broad or even effective. “If you can accomplish the same goal some other way, or if this law doesn’t accomplish the goal, then the court will say, ‘We’re going to strike it down. Sure this is a good, worthy goal, but this law just doesn’t advance it.’” Sacharoff says. “So it will be very fact specific.”

1 0

What a report from Germany teaches us about investigating algorithms - Columbia Journalism Review
by J.C. DE MARTIN Jan. 11, 2019

Jan. 11, 2019

https://www.cjr.org/tow_center/investigating-algorithims-germany-schufa.php (Sent from my wireless device; please excuse brevity and typos (if any))

3 4