336 Matching Annotations
  1. Last 7 days
    1. Building on platforms' stores of user-generated content, competing middleware services could offer feeds curated according to alternate ranking, labeling, or content-moderation rules.

      Already I can see too many companies relying on artificial intelligence to sort and filter this material and it has the ability to cause even worse nth degree level problems.

      Allowing the end user to easily control the content curation and filtering will be absolutely necessary, and even then, customer desire to do this will likely loose out to the automaticity of AI. Customer laziness will likely win the day on this, so the design around it must be robust.

  2. Jul 2021
    1. Facebook AI. (2021, July 16). We’ve built and open-sourced BlenderBot 2.0, the first #chatbot that can store and access long-term memory, search the internet for timely information, and converse intelligently on nearly any topic. It’s a significant advancement in conversational AI. https://t.co/H17Dk6m1Vx https://t.co/0BC5oQMEck [Tweet]. @facebookai. https://twitter.com/facebookai/status/1416029884179271684

  3. Jun 2021
    1. intelligence collective réflexive

      Il ne s’agirait donc pas de simplement devenir collectivement «plus intelligent» (au sens d’efficace, dans un strict paradigme scientifique et technique pour accélérer le fonctionnement de l’économie), mais aussi réflexif: réfléchir aux conditions de cette société renouvelée, en proie à de nouvelles dynamiques de pouvoir extrêmement concentrées et asymétriques.

    1. t hadn’t learned sort of the concept of a paddle or the concept of a ball. It only learned about patterns of pixels.

      Cognition and perception are closely related in humans, as the theory of embodied cognition has shown. But until the concept of embodied cognition gained traction, we had developed a pretty intellectual concept of cognition: as something located in our brains, drained of emotions, utterly rational, deterministic, logical, and so on. This is still the concept of intelligence that rules research in AI.

    2. the original goal at least, was to have a machine that could be like a human, in that the machine could do many tasks and could learn something in one domain, like if I learned how to play checkers maybe that would help me learn better how to play chess or other similar games, or even that I could use things that I’d learned in chess in other areas of life, that we sort of have this ability to generalize the things that we know or the things that we’ve learned and apply it to many different kinds of situations. But this is something that’s eluded AI systems for its entire history.

      The truth is we do not need to have computers to excel in the things we do best, but to complement us. We shall bet on cognitive extension instead of trying to re-create human intelligence --which is a legitimate area of research, but computer scientists should leave this to cognitive science and neuroscience.

    1. Last year, Page told a convention of scientists that Google is “really trying to build artificial intelligence and to do it on a large scale.”

      What if they're not? What if they're building an advertising machine to manipulate us into giving them all our money?

      From an investor perspective, the artificial answer certainly seems sexy while using some clever legerdemain to keep the public from seeing what's really going on behind the curtain?

    2. It seeks to develop “the perfect search engine,” which it defines as something that “understands exactly what you mean and gives you back exactly what you want.”

      What if we want more serendipity? What if we don't know what we really want? Where is this in their system?

  4. May 2021
    1. Turing was an exceptional mathematician with a peculiar and fascinating personality and yet he remains largely unknown. In fact, he might be considered the father of the von Neumann architecture computer and the pioneer of Artificial Intelligence. And all thanks to his machines; both those that Church called “Turing machines” and the a-, c-, o-, unorganized- and p-machines, which gave rise to evolutionary computations and genetic programming as well as connectionism and learning. This paper looks at all of these and at why he is such an often overlooked and misunderstood figure.
  5. Mar 2021
    1. In this respect, we join Fitzpatrick (2011) in exploring “the extent to which the means of media production and distribution are undergoing a process of radical democratization in the Web 2.0 era, and a desire to test the limits of that democratization”

      Something about this is reminiscent of WordPress' mission to democratize publishing. We can also compare it to Facebook whose (stated) mission is to connect people, while it's actual mission is to make money by seemingly radicalizing people to the extremes of our political spectrum.

      This highlights the fact that while many may look at content moderation on platforms like Facebook as removing their voices or deplatforming them in the case of people like Donald J. Trump or Alex Jones as an anti-democratic move. In fact it is not. Because of Facebooks active move to accelerate extreme ideas by pushing them algorithmically, they are actively be un-democratic. Democratic behavior on Facebook would look like one voice, one account and reach only commensurate with that person's standing in real life. Instead, the algorithmic timeline gives far outsized influence and reach to some of the most extreme voices on the platform. This is patently un-democratic.

    1. Meanwhile, the algorithms that recommend this content still work to maximize engagement. This means every toxic post that escapes the content-moderation filters will continue to be pushed higher up the news feed and promoted to reach a larger audience.

      This and the prior note are also underpinned by the fact that only 10% of people are going to be responsible for the majority of posts, so if you can filter out the velocity that accrues to these people, you can effectively dampen down the crazy.

    2. In his New York Times profile, Schroepfer named these limitations of the company’s content-moderation strategy. “Every time Mr. Schroepfer and his more than 150 engineering specialists create A.I. solutions that flag and squelch noxious material, new and dubious posts that the A.I. systems have never seen before pop up—and are thus not caught,” wrote the Times. “It’s never going to go to zero,” Schroepfer told the publication.

      The one thing many of these types of noxious content WILL have in common are the people at the fringes who are regularly promoting it. Why not latch onto that as a means of filtering?

    3. But anything that reduced engagement, even for reasons such as not exacerbating someone’s depression, led to a lot of hemming and hawing among leadership. With their performance reviews and salaries tied to the successful completion of projects, employees quickly learned to drop those that received pushback and continue working on those dictated from the top down.

      If the company can't help regulate itself using some sort of moral compass, it's imperative that government or other outside regulators should.

    4. <small><cite class='h-cite via'> <span class='p-author h-card'>Joan Donovan, PhD</span> in "This is just some of the best back story I’ve ever read. Facebooks web of influence unravels when @_KarenHao pulls the wrong thread. Sike!! (Only the Boston folks will get that.)" / Twitter (<time class='dt-published'>03/14/2021 12:10:09</time>)</cite></small>

    1. System architects: equivalents to architecture and planning for a world of knowledge and data Both government and business need new skills to do this work well. At present the capabilities described in this paper are divided up. Parts sit within data teams; others in knowledge management, product development, research, policy analysis or strategy teams, or in the various professions dotted around government, from economists to statisticians. In governments, for example, the main emphasis of digital teams in recent years has been very much on service design and delivery, not intelligence. This may be one reason why some aspects of government intelligence appear to have declined in recent years – notably the organisation of memory.57 What we need is a skill set analogous to architects. Good architects learn to think in multiple ways – combining engineering, aesthetics, attention to place and politics. Their work necessitates linking awareness of building materials, planning contexts, psychology and design. Architecture sits alongside urban planning which was also created as an integrative discipline, combining awareness of physical design with finance, strategy and law. So we have two very well-developed integrative skills for the material world. But there is very little comparable for the intangibles of data, knowledge and intelligence. What’s needed now is a profession with skills straddling engineering, data and social science – who are adept at understanding, designing and improving intelligent systems that are transparent and self-aware58. Some should also specialise in processes that engage stakeholders in the task of systems mapping and design, and make the most of collective intelligence. As with architecture and urban planning supply and demand need to evolve in tandem, with governments and other funders seeking to recruit ‘systems architects’ or ‘intelligence architects’ while universities put in place new courses to develop them.
  6. Feb 2021
  7. Jan 2021
    1. As an opening move, I’d suggest that we could reconceptualize intelligence as NaQ (neuroacoustic quotient), or ‘the capacity to cleanly switch between different complex neuroacoustic profiles.’

      also seems more neutral and embracing the differences in [[neurodiversity]] / individual thinking vs relentless optimizing for a certain KPI (like for IQs) #[[to write]]

  8. Dec 2020
    1. création collective de sens qui est au cœur de l’intelligence humaine

      objectif de l'intelligence collective, des humanités numériques comme discipline en communauté

    2. les chercheurs en sciences humaines doivent donner l’exemple – dans leur pratique ! – d’une production de sens qui s’offre à la connaissance de la manière la plus transparente possible

      Injonction aux faiseurs de connaissance – autre morceau du programme de l’intelligence collective de Pierre Lévy?

      versant éthique?

  9. Nov 2020
  10. Oct 2020
    1. Australia's Cyber Security Strategy: $1.66 billion dollar cyber security package = AFP gets $88 million; $66 million to critical infrastructure organisations to assess their networks for vulnerabilities; ASD $1.35 billion (over a decade) to recruit 500 officers.

      Reasons Dutton gives for package:

      • child exploitation
      • criminals scamming, ransomware
      • foreign governments taking health data and potential attacks to critical infrastructure

      What is defined as critical infrastructure is expanded and subject to obligations to improve their defences.

      Supporting cyber resilience of SMEs through information, training, and services to make them more secure.

    1. Similarly, technology can help us control the climate, make AI safe, and improve privacy.

      regulation needs to surround the technology that will help with these things

    1. What if you could use AI to control the content in your feed? Dialing up or down whatever is most useful to you. If I’m on a budget, maybe I don’t want to see photos of friends on extravagant vacations. Or, if I’m trying to pay more attention to my health, encourage me with lots of salads and exercise photos. If I recently broke up with somebody, happy couple photos probably aren’t going to help in the healing process. Why can’t I have control over it all, without having to unfollow anyone. Or, opening endless accounts to separate feeds by topic. And if I want to risk seeing everything, or spend a week replacing my usual feed with images from a different culture, country, or belief system, couldn’t I do that, too? 

      Some great blue sky ideas here.

    1. Walter Pitts was pivotal in establishing the revolutionary notion of the brain as a computer, which was seminal in the development of computer design, cybernetics, artificial intelligence, and theoretical neuroscience. He was also a participant in a large number of key advances in 20th-century science.
  11. Sep 2020
    1. doivent consulter des oracles

      Par exemple, pour entraîner une intelligence artificielle, le philosophe montréalais Martin Gibert propose de montrer aux IA des exemples, des modèles à suivre (des Greta et des Mère Theresa) plutôt que d’essayer de leur enseigner les concepts de la philosophie morale.

  12. Aug 2020
    1. Advantages of people in [[Silicon Valley]]:** super smart but not necessarily highly educated so they don’t just believe what everyone else does. **They think outside the box. They’re thinkers as well as people that have had to do things and pass [[reality]] tests. The only test most academics face is "can I publish this piece?"

      What differs people in Silicon Valley and typical students

  13. Jul 2020
  14. Jun 2020
    1. But tagging, alone, is still not good enough. Even our many tags become useless if/when their meaning changes (in our minds) by the time we go retrieve the data they point to. This could be years after we tagged something. Somehow, whether manually or automatically, we need agents and tools to help us keep our tags updated and relevant.

      search engines usually can surface that faster (less cognitive load than recalling what and where you store something) than you retrieve it in your second brain (abundance info, do can always retrieve from external source in a JIT fashion)

    1. each of them flows through each of the two layers of the encoder

      each of them flows through each of the two layers of EACH encoder, right?

    1. It made it challenging for the models to deal with long sentences.

      This is similar to autoencoders struggling with producing high-resolution imagery because of the compression that happens in the latent space, right?

    1. it seems that word-level models work better than character-level models

      Interesting, if you think about it, both when we as humans read and write, we think in terms of words or even phrases, rather than characters. Unless we're unsure how to spell something, the characters are a secondary thought. I wonder if this is at all related to the fact that word-level models seem to work better than character-level models.

    2. As you can see above, sometimes the model tries to generate latex diagrams, but clearly it hasn’t really figured them out.

      I don't think anyone has figured latex diagrams (tikz) out :')

    3. Antichrist

      uhhh should we be worried

    1. We only forget when we’re going to input something in its place. We only input new values to the state when we forget something older.

      seems like a decision aiming for efficiency

    2. outputs a number between 000 and 111 for each number in the cell state Ct−1Ct−1C_{t-1}

      remember, each line represents a vector.

    1. Just as journalists should be able to write about anything they want, comedians should be able to do the same and tell jokes about anything they please

      where's the line though? every output generates a feedback loop with the hivemind, turning into input to ourselves with our cracking, overwhelmed, filters

      it's unrealistic to wish everyone to see jokes are jokes, to rely on journalists to generate unbiased facts, and politicians as self serving leeches, err that's my bias speaking

  15. May 2020
    1. Mei, X., Lee, H.-C., Diao, K., Huang, M., Lin, B., Liu, C., Xie, Z., Ma, Y., Robson, P. M., Chung, M., Bernheim, A., Mani, V., Calcagno, C., Li, K., Li, S., Shan, H., Lv, J., Zhao, T., Xia, J., … Yang, Y. (2020). Artificial intelligence for rapid identification of the coronavirus disease 2019 (COVID-19). MedRxiv, 2020.04.12.20062661. https://doi.org/10.1101/2020.04.12.20062661

    1. Shweta, F., Murugadoss, K., Awasthi, S., Venkatakrishnan, A., Puranik, A., Kang, M., Pickering, B. W., O’Horo, J. C., Bauer, P. R., Razonable, R. R., Vergidis, P., Temesgen, Z., Rizza, S., Mahmood, M., Wilson, W. R., Challener, D., Anand, P., Liebers, M., Doctor, Z., … Badley, A. D. (2020). Augmented Curation of Unstructured Clinical Notes from a Massive EHR System Reveals Specific Phenotypic Signature of Impending COVID-19 Diagnosis [Preprint]. Infectious Diseases (except HIV/AIDS). https://doi.org/10.1101/2020.04.19.20067660

  16. Apr 2020
    1. Abdulla, A., Wang, B., Qian, F., Kee, T., Blasiak, A., Ong, Y. H., Hooi, L., Parekh, F., Soriano, R., Olinger, G. G., Keppo, J., Hardesty, C. L., Chow, E. K., Ho, D., & Ding, X. (n.d.). Project IDentif.AI: Harnessing Artificial Intelligence to Rapidly Optimize Combination Therapy Development for Infectious Disease Intervention. Advanced Therapeutics, n/a(n/a), 2000034. https://doi.org/10.1002/adtp.202000034

    1. The world’s largest exhibitions organizer, London-based Informa plc, outlined on Thursday morning a series of emergency actions it’s taking to alleviate the impact of the COVID-19 pandemic on its events business, which drives nearly two-thirds of the company’s overall revenues. Noting that the effects have been “significantly deeper, more volatile and wide-reaching,” than was initially anticipated, the company says it’s temporarily suspending dividends, cutting executive pay and issuing new shares worth about 20% of its total existing capital in an effort to strengthen its balance sheet and reduce its approximately £2.4 billion ($2.9 billion) in debt to £1.4 billion ($1.7 billion). Further, Informa says it’s engaged in “constructive discussions” with its U.S.-based debt holders over a covenant waiver agreement.

      Informa Group, que posee editoriales como Taylor & Francis, de Informa Intelligent Division toma medidas en su sector de conferencias y eventos. Provee dos tercios de sus ingresos totales, 2.9 billion dólares. Emite acciones y para el mercado norteamericano acuerdos de deuda. Mientras la parte editorial que aporta un 35% de los ingresos se mantiene sin cambios y con pronósticos estables y sólidos. Stephen Carter CEO

    1. Le public acquiert ainsi une nouvelle fonction : celle d’instance critique auquel doit s’exposer le pouvoir.

      fonction de l’espace public: un appareil critique (la critique est productrice d’espace public).

      le pouvoir, pour maintenir sa légitimité, doit être exposé à la sphère publique et se montrer à lui avec transparence; il doit pouvoir être challengé; s’il ne résiste pas à la critique publique, il ne mérite pas d’être en place.

      la possibilité de challenger l’instance publique est comparable à la publication des protocoles de sécurité utilisées dans le domaine public (ex. SSL/TLS): la sécurité des contenus encryptés tirent justement leur robustesse du fait que leur algorithme est public; quiconque pourrait le challenger à tout moment, si bien qu’on s’assure d’en éliminer toutes les failles (et l’intelligence collective peut être mise à contribution, le cas échéant).

    1. Des applications de visites guidées intelligentes s’appuient sur un processus de gestion des flux visiteurs (Visitor Flow Management Process, VFMP) pour les orienter vers les zones où ils sont le moins nombreux. Il s’agira alors de combiner les données sur l’affluence en temps réel pour chaque espace avec les souhaits et les goûts des visiteurs pour suggérer le parcours personnalisé idéal

      Argument en faveur de l'IA qui permet bien de gérer le flux mais ajoute un second bénéfice : proposer un parcours idéal. Ce bénéfice supplémentaire peut être considéré comme un argument réthorique de type Logos.

    2. Certaines technologies intelligentes utilisées dans d’autres secteurs pourraient être transposées dans les musées. Avec le big data, il est possible de connaître l’affluence en fonction des dates et des horaires, les types de visiteurs selon les jours et les périodes, ou la durée de visite moyenne par rapport différents paramètres comme la météo.

      Argument épistémique inductif et réthorique de type logos.

      On passe à l'intelligence artificielle, technologie de pointe. Apporte du crédit à l'affirmation du bénéfice du numérique.

  17. Feb 2020
    1. visual are processed 60,000 times faster in the brain than text and visual aids in the classroom improve learning up to 400 percent. Ideas presented graphically are easier to understand and remember than those presented as words, (Kliegel et al., 1987).

      throw out this factoid when doing video?

  18. Dec 2019
    1. Ranking the intelligence of animals seems an increasingly pointless exercise when one considers the really important thing: how well that animal is adapted to its niche
    1. “NextNow Collaboratory is an interesting example of a new kind of collective intelligence: an Internet-enabled, portable social network, easily transferable from one social cause to another.”

      Sense Collective's TotemSDK brings together tools, protocols, platform integrations and best practices for extending collective intelligence beyond our current capabilities. A number of cryptographic primitives have emerged which support the amazing work of projects like the NextNow Collaboratory in exciting ways that help to upgrade the general purpose social computing substrate which make tools like hypothes.is so valuable.

    1. A natural language provides its user with a ready-made structure of concepts that establishes a basic mental structure, and that allows relatively flexible, general-purpose concept structuring. Our concept of language as one of the basic means for augmenting the human intellect embraces all of the concept structuring which the human may make use of.
    2. It has been jokingly suggested several times during the course of this study that what we are seeking is an "intelligence amplifier." (The term is attributed originally to W. Ross Ashby[2,3]. At first this term was rejected on the grounds that in our view one's only hope was to make a better match between existing human intelligence and the problems to be tackled, rather than in making man more intelligent. But deriving the concepts brought out in the preceding section has shown us that indeed this term does seem applicable to our objective. 2c2a Accepting the term "intelligence amplification" does not imply any attempt to increase native human intelligence. The term "intelligence amplification" seems applicable to our goal of augmenting the human intellect in that the entity to be produced will exhibit more of what can be called intelligence than an unaided human could; we will have amplified the intelligence of the human by organizing his intellectual capabilities into higher levels of synergistic structuring. What possesses the amplified intelligence is the resulting H-LAM/T system, in which the LAM/T augmentation means represent the amplifier of the human's intelligence.2c2b In amplifying our intelligence, we are applying the principle of synergistic structuring that was followed by natural evolution in developing the basic human capabilities. What we have done in the development of our augmentation means is to construct a superstructure that is a synthetic extension of the natural structure upon which it is built. In a very real sense, as represented by the steady evolution of our augmentation means, the development of "artificial intelligence" has been going on for centuries.
    1. This is not a new idea. It is based on the vision expounded by Vannevar Bush in his 1945 essay “As We May Think,” which conjured up a “memex” machine that would remember and connect information for us mere mortals. The concept was refined in the early 1960s by the Internet pioneer J. C. R. Licklider, who wrote a paper titled “Man-Computer Symbiosis,” and the computer designer Douglas Engelbart, who wrote “Augmenting Human Intellect.” They often found themselves in opposition to their colleagues, like Marvin Minsky and John McCarthy, who stressed the goal of pursuing artificial intelligence machines that left humans out of the loop.

      Seymour Papert, had an approach that provides a nice synthesis between these two camps, buy leveraging early childhood development to provide insights on the creation of AI.

    2. Thompson’s point is that “artificial intelligence” — defined as machines that can think on their own just like or better than humans — is not yet (and may never be) as powerful as “intelligence amplification,” the symbiotic smarts that occur when human cognition is augmented by a close interaction with computers.

      Intelligence amplification over artificial intelligence. In reality you can't get to AI until you've mastered IA.