Categories
AI

Comment: DeepMind Can Now Beat Us at Multiplayer Games, Too

DeepMind’s agents are not really collaborating, said Mark Riedl, a professor at Georgia Tech College of Computing who specializes in artificial intelligence. They are merely responding to what is happening in the game, rather than trading messages with one another, as human players do…Although the result looks like collaboration, the agents achieve it because, individually, they so completely understand what is happening in the game.

Metz, C. (2019). DeepMind Can Now Beat Us at Multiplayer Games, Too. New York Times.

The problem with arguments like this is that 1) we end up playing semantic games about what words mean, 2) what we call the computer’s achievement isn’t relevant, and 3) just because the algorithmic solution doesn’t look the same as a human solution doesn’t make it less effective.

The concern around the first point is that, as algorithms become more adept at solving complex problems, we end up painting ourselves into smaller and smaller corners, hemmed in by how we defined the characteristics necessary to solve those problems. In this case, we can define collaboration in a way that means that algorithms aren’t really collaborating but tomorrow when they can collaborate according to today’s definition, we’ll see people wanting to change the definition again.

The second point relates to competence. Algorithms are designed to be competent at solving complex problems, not to solve them in ways that align with our definitions of what words mean. In other words, DeepMind doesn’t care how the algorithm solves the problem, only that it does. Think about developing a treatment for cancer…will we care that the algorithm didn’t work closely with all stakeholders, as human teams would have to, or will it only matter that we have an effective treatment? In the context of solving complex problems, we care about competence.

And finally, why would it matter that algorithmic solutions don’t look the same as human solutions? In this case, human game-players have to communicate in order to work together because it’s impossible for them to do the computation necessary to “completely understand what is happening in the game”. If we had the ability to do that computation, we’d also drop “communication” requirement because it would only slow us down and add nothing to our ability to solve the problem.

Categories
education learning

Posted to Diigo 11/18/2011

    • The really basic skill today is the skill of learning, and the best use of games is to leverage their tendency to enhance it
    • talking about games and learning is an important activity
    • when they get the support and have access to suitable software systems, children’s enthusiasm for playing games easily gives rise to an enthusiasm for making them, and this in turn leads to more sophisticated thinking about all aspects of games
    • the idea that children should draw, write stories and play music is not contradicted by the fact that their work is not of professional quality

Posted from Diigo. The rest of my favorite links are here.

Categories
twitter feed

Twitter Weekly Updates for 2011-11-07

Categories
twitter feed

Twitter Weekly Updates for 2011-08-29

  • @AMEE_Online this is great, how do we go about claiming the year’s membership? #
  • RT @Jane_Mooney: Great game-based learning resources for educators from @judithway http://t.co/f1wyv1P #
  • Just registered for #amee2011 after spending 19 hours in transit. The world is smaller than it used to be but it could be smaller still #
  • @jane_mooney I’ll look out for u & your poster. If u want 2 chat I’d love to hook up. My PhD is on blended learning in clinical education #
  • @paulderoos Good luck with the free accommodation for #amee2011 I’d help you out if I could 🙂 #
  • Gearing up for #amee2011 where I’ll b presenting a systematic review on blended learning in clinical education. Let me know if u’ll b there #
  • @amcunningham Official AMEE & Medical Teacher twitter accounts are using #amee2011 #
  • @amcunningham no doubt there are good sessions, it’s just all a bit overwhelming right now. Trying to make some sense of the programme #
  • @amcunningham nothing official about hashtags, just assumed it’d be the full date, will use whatever the convention is 🙂 #
  • @amcunningham I’ve been looking at the presentation sessions for that period & nothing has grabbed me yet. Maybe I’ll come to your workshop #
  • @amcunningham Yes, I’ll be at #amee2011 starting to get excited about it now. We’ll definitely hook up 🙂 #
  • @amcunningham See you’re facilitating a workshop on social media at #amee2011 You know what level the session is aimed at? #
  • Announcing the Zotero 3.0 Beta Release http://t.co/EmDnU32 #
Categories
twitter feed

Twitter Weekly Updates for 2011-04-18

Categories
twitter feed

Twitter Weekly Updates for 2011-02-28

  • @hotdogcop Tweeted this earlier, pretty funny, but also pretty accurate http://onion.com/h3coIN #
  • The Changing Landscape of Higher Education (EDUCAUSE Review) | EDUCAUSE http://bit.ly/dFXmaz #
  • Richard Feynman on the pleasure of finding things out http://bit.ly/gij2ve #
  • Massive Health Uses Big Data, Mobile Phones to Fight Chronic Disease http://ow.ly/1s4Epd #
  • The Daily Is Interesting, But Is It the Future of Newspapers? http://ow.ly/1s4En9 #
  • Gladwell Still Missing the Point About Social Media and Activism http://ow.ly/1s4ElH #
  • “This Game Sucks”: How to Improve the Gamification of Education (EDUCAUSE Review) | EDUCAUSE http://bit.ly/fMWN4f #
  • eLearn: Feature Article – The Effects of Twitter in an Online Learning Environment http://bit.ly/gqrqfQ. Students resist adoption of Twitter #
  • eLearn: Feature Article – Administering a Gross Anatomy Exam Using Mobile Technology http://bit.ly/f74yAj #
  • Onion Report: Increasing Number Of Educators Found To Be Suffering From Teaching Disabilities http://onion.com/h3coIN. Humour #
  • Skateboarding Physics Professor At Large — Blog Archive » Building A New Culture Of Teaching And Learning http://bit.ly/hco11U #
  • Presentation Zen: Nurturing curiosity & inspiring the pursuit of discovery http://bit.ly/h4SJtd. Why don’t we value curiosity in schools? #
  • @hotdogcop Seems to be this idea that “if you build it, they will come”, but the reality is that no-one knows what to do when they get there #
  • The Need for Student Social Media Policies (EDUCAUSE Review) | EDUCAUSE http://bit.ly/fhj7F9 #
  • Evidence of Learning Online: Assessment Beyond The Paper — Campus Technology http://bit.ly/e8iuuQ #
  • E10 Podcast: Gardner Campbell and Jim Groom Discuss Faculty Attitudes and the Joy of Learning | EDUCAUSE http://bit.ly/fmLBnk #
  • The Daily Papert. Words and wisdom of Dr. Seymour Papert http://bit.ly/fs947e #
  • Mendeley Update: OpenURL support, improved article pages, & easier citation entry for OpenOffice http://bit.ly/fyOwcc #
  • Thought Leader » John Vlismas » Hofmeyr, Bloody Hofmeyr http://bit.ly/eNQumA. Intelligent response to Steve’s rant #
  • Managing your research the modern way: Research together with colleagues using an activity feed | Mendeley Blog http://bit.ly/e1eQAS #
  • The Enormous Technological Challenges Facing Education http://bit.ly/eVBik5. Nice summary of the emerging tech in the latest Horizon report #
  • A WikiLeaks Clone Takes On Higher Education – Wired Campus – The Chronicle of Higher Education http://bit.ly/g5WL4W #
  • @subcide Great, thanks for the update. #Mendeley is one of the tools I use most often and it’s brilliant to see it continually improving #
  • Thought leader: nothing to correct http://ht.ly/40Xq0. The problem of “corrective rape” in SA, aimed at the LGBTI community. I wasn’t aware #
  • The ‘myth’ of e-learning http://bit.ly/gXQmUx #
  • @dgachago17 #Hootsuite (web app) is also pretty good at updating multiple services simultaneously #justsaying #
  • Not sure which version of #Mendeley added this, but I just found the “Send by email” feature, and it is SOOO welcome 🙂 #
  • Acceptable Use Policies in a Web 2.0 & Mobile Era: A Guide for School District ~ Stephen’s Web http://bit.ly/i1MBPZ #
  • Critical Thinking: More than Words? http://bit.ly/i5rLtO. We talk about critical thinking with our students, but don’t discuss what it means #
Categories
diigo

Posted to Diigo 05/16/2010

    • “PLEs are more a methodology or an approach to technology enhanced learning than an application.”
    • the introduction of e-learning led to a reverse in pedagogic innovation
    • I tend to think that knowledge is best shared and developed through communities of practice. Communities of practice as Etienne Wenger says are based on a shared repertoire of communal resources (routines, sensibilities, artefacts, vocabulary, styles, etc.) that members have developed over time.
    • One of the challenges faced by traditional education is found in the very goal of its existence (second only to its role as providing value statements through accreditation): to present bounded structures of knowledge in order for others to learn a discipline. Or put another way, schools and universities help students make sense of a discipline.
    • There are two elements under consideration:
    • 1. The curriculum itself
    • This is what learners must learrn.
    • Curriculum/content is created and disseminated through research and publication. This content then forms the basis of instruction. Nothing new here, with the exception of the argument that the scholarly publication process is too slow.
    • It’s this content that most people see as the important part of education.
    • We make learners do all sorts of fun things to get this to happen: cases, problem-based learning, games/simulations, lectures, podcasts, tests, eportfolios, and so on.
    • 2. The framework of sensemaking
    • Should the educator provide a formed narrative of coherence? Or should learners be tasked with this? Should the educator create a fully bounded content structure? Or should the content interaction opportunities be more fluid? And what about interaction? Should it be under the control of educators? Or should learners self-organanize as they deem worthwhile? This is where education truly begins to change. Tweaking content creating and delivery models is perhaps a start. But it’s not transformative.
    • It means the conversation is more chaotic. It means that we’re always missing something. Everyone is. Some important conversation, somewhere, is being overlooked. Why is that so discomforting?
    • We expect the academy to be a place that provide clarity, a path forward
    • When we then step into a course and discover the conversation is distributed and that the expected frameworks for telling us what to think don’t exist, we get disoriented.

      But isnt’ that life?

    • Isn’t that how real learning occurs? In business? In our personal lives? Who actually possesses a framework fo sensemaking in advance of encountering novel problems? Or who can rely on the “narrative of coherence” provided in advance of becoming a parent?
    • We can’t manage it all. We must choose. As we move through this course, we’ll focus more on what it means to choose – i.e. what types of networks we want and need to build. For now, realizing that our ability to make sense is under our control.
    • Where is the learning in this? The learning exists in the process of forming and navigating networks. Some sources we filter. Some thinkers we value greatly. Others we ignore.
    • This paper described experimentation in the development of distributed online courses and in software – particularly, the personal learning environment – that support the formation of connections between the far-flung pieces of such courses
    • suggests a pedagogy of participation rather than retention, and even suggests distributed and locally-based forms of evaluation and assessment
    • The intent of such systems is to to facilitate the conversation and interaction around episodic learning events in a distributed environment, transforming them from elements in a linear flow-based design to free-floating objects in an environment
    • In addition to providing an engaging and immersive environment for student learning, substantially improving motivation and interaction with the learning material, games and simulations are able to support learning in complex environments, offering a subtlety simple instruction-based or lecture-based learning cannot offer. (Squire, 2005)
    • games and simulations fall into a category similar to lectures and presentations in that they involve statically designed learning objectives and strategies. (Amory & Seagram, 2003)
    • we often do not know what it is we want to teach the student. (Caine & Caine, 1997) Today’s environment is variable, which means situations – and hence, fact – change fluidly. One day Pluto is a planet, the next day it is not. One day Czechoslovakia is a country, the next day it is not. One day capitalism is the unassailable foundation for our economic system, the next day, following a market collapse, it is not. Moreover, today’s environment is complex. The relations between variables cannot be described or even predicted.
    • learners themselves are changing
    • It has even been suggested that our interactions with modern communication technologies change the way we think. Even if we reject such descriptions as students as overly broad and inaccurate generalizations – and there is good reason for doing so – it is nonetheless the case that the needs, capabilities and interests of the target audience is rapidly shifting and changing
    • It is not merely to create a network into which to situate episodic learning, but rather, to create a network that learns and thus adapts and reshapes itself based on those conversations and interactions. (Downes, 2007)
    • the best we can manage is to teach students how to learn, and to encourage them to manage their own learning thereafter
    • how we learn itself is something that changes, and cannot be precisely taught
    • For this reason, we need to see the educational system itself as adaptive rather than merely prescriptive
    • knowledge exists in the minds of the members or participants, and this knowledge is derived from their direct (and recent) experience in the field
    • In addition, the need for content and support emerges from conversations among the participants. These interactions are able to reveal not only what company commanders know, but also what they don’t know (and need to know). The interaction, in other words, meets and addresses an objection often put of self-directed learners, that they don’t know what they need to know. (Clayson, 2005)
    • The core of a social networking technology is the capacity to create links between members in a community – to create, in other words, social networks
    • Topics, for example, are not assigned centrally, but are instead created by individuals ‘tagging’ certain content with terms or categories they choose themselves. (Barsky & Purdon, 2006) Each person’s social network on a social networking site, moreover, is unique; there is no definitive grouping of people, only a clustering of people with more or less similar interests.
    • users jump from service to service, creating (and discarding) new identities as needed. A typical web user may have multiple ‘home pages’ – their personal blog, their photo page on Flickr or Picassa, their Google Reader account, shared documents through Zoho, their video page on YouTube, their Twitter account, their profiles, on Facebook, MySpace and LinkedIn, their Wikipedia login, their email accounts, and (often least) their university LMS login. While real friendships and communities develop through this mélange, loyalty to online sites and services is limited and fleeting. (O’Brien, 2007)
    • The idea of the personal learning environment is that it performs many of the functions of a content management system and of a social network system but from the perspective of the individual rather than the community or the institution. (Attwell, 2006) Hence, the PLE may be understood as the intersection of the multiple home pages employed by any given individual. In the first instance, the PLE is a concept, rather than an application – it is the idea that a person’s web presence can be distributed. (Attwell, Graham Attwell: “Knowledge is best shared and developed through communities of practice”, 2007)
    • Because there were so many people contributing to the course, and because the content of the course actually shifted and varied according to participation and input into the course, it was necessary to emphasize to students that their role in the course was not to attempt to assimilate all course content. This was neither possible nor desirable. Rather, students were told that their role was to select and sample course content, pursuing areas of interest, reading related material from both within and outside the course, and then to contribute their unique perspective based on this reading. (Siemens, Where does the learning occur??, 2008)
    • we are currently seeing experimentation in the development of distributed online courses and in software – particularly, the personal learning environment – that support the formation of connections between the far-flung pieces of such courses
    • In the PLE project being undertaken by the National Research Council, the functionality of the PLE is depicted in four major stages: to aggregate, that is, to collect content from the individual’s and other online content service providers, where aggregation includes elements of recommendation, data mining and automated metadata extraction ; to remix, or to organize content from several different sources in different ways, including through automated clustering; to repurpose, or edit, localize, or otherwise modify or create new content; and to feed forward, or send the content to subscribers and other web services, either via RSS syndication, email, Twitter, or other relevant services. (Downes, Theory of Learning Networks, 2004)
    • the PLE is not a recreation of the capabilities of the learning management system, but rather, a learning network