Computing At School

---

Resources for CAS-AI Discussion List
Making Artificial Intelligence Programming available in Schools,
for teaching AI as Science

---

John McCarthy, one of the Founders of AI, who coined the name "Artificial Intelligence", died in October 2011, aged 84. For him, AI was primarily science, not engineering. http://www-formal.stanford.edu/jmc/ Image source http://www-formal.stanford.edu/jmc/personal.html A personal memoir.

---

This is part of The Free Poplog Portal
Liberal copyright conditions for everything on this web site

---

DISCLAIMER

THIS IS NOT AN OFFICIAL DOCUMENT OF THE UNIVERSITY OF BIRMINGHAM OR THE SCHOOL OF COMPUTER SCIENCE.
NEITHER THE UNIVERSITY NOR THE SCHOOL HAS ENDORSED THE OPINIONS EXPRESSED HERE.

---

PARTIAL, PROVISIONAL, DESCRIPTION OF THE CAS-AI GROUP
FOR TEACHERS INTERESTED IN TEACHING AI AS SCIENCE

Programming is much more than composing collections of instructions. It includes design, implementation, testing, debugging, analysis, criticism, comparison, and extension of working programs. This broad and deep collection of competences should be included in school curricula not just because industry needs programmers, but also because many of the major SCIENTIFIC advances of future decades will depend on researchers and practitioners of many kinds including scientists, educators, philosophers and law-makers, knowing how to investigate, think about, model, and explain complex natural information-processing systems, and being able to devise useful ways of helping, repairing or supporting them when things go wrong.

• Introduction
  • Purpose of the CAS-AI group
  • Thinky programming for young learners
  • The cas-ai mailing list
  • Form for list joiners
• Materials related to AI programming in Pop11 (an AI language that grows with the user)
  • Materials for beginners (including teachers)
    • Some experimental Video tutorials (podcasts?) on AI concepts and techniques
    • Getting started with the Ved/XVed editor.
    • Some introductory AI-ish programming exercises (text only)
    • Pop11 Syntax (an introductory subset)
    • Examples using graphics
  • Web site with introductions to "thinky programming" using Pop11
  • A recorded 'Teach share' presentation
  • New version of POP11 PRIMER (Version 4)
• Proposal for teaching AI/Cognitive Science at A-Level AS-level
• AI Toolkit and teaching materials on Amazon Cloud
• Learning AI on a remote Linux machine.
• Useful links for people not familiar with linux/unix.
• Other AI Resources

---

Introduction

• Purpose of the CAS-AI group
  The decision to create this group, including the mailing list and associated resources, arose out of email discussions among members of CAS (ComputingAtSchool) http://www.computingatschool.org.uk/ about the need for education in computing to go beyond the normal "utilitarian" view of computing, which values programming only as a means to producing useful and entertaining applications, contributing to the economy, meeting the needs of employers, providing entertainment, providing challenging jobs, etc.

  In particular, there is also a need to contribute to scientific and philosophical education so that some young learners, especially future scientists, learn how construction of working programs can increase our scientific understanding of some complex natural systems, by offering models or explanations of various aspects of information processing in humans and other animals (not only individual animals but also swarms, flocks, social and economic systems). For example, learning about software virtual machines running in computers can provide new ideas about how minds might be related to brains.

  We could update the ancient Greek maxim "Know thyself" to:

      "Learn programming and begin to model yourself".
  

  AI research on learning, seeing, planning, controlling actions, understanding and generating natural language, modelling reasoning processes, modelling mathematical discovery, modelling creative painting or musical composition (among others) can provide new ideas about how human and animal minds work, and can inspire kinds of research on what brains do.

  The CAS-AI group provides

  • an email discussion list,
  • access to an AI teaching and learning environment hosted on an Amazon virtual machine.
  • pointers to versions that can be used locally on linux and windows machines, and a reduced version, without graphics for use on windows (see Option 2).
  • Other educational AI resources for schools (To be Added -- please submit suggestions)
    • See the 'Udacity' web site offering online AI courses.
    • CAS members who are registered for Udacity courses and wish to join disussions of those courses are invited to subscribe to a forum at goo.gl/n6twp.

• Thinky programming for young learners
  For this initiative to succeed, we need at least some schools to introduce ideas, techniques, programming languages and tools related to the development of "thinky" programs, in the sense defined here: http://www.cs.bham.ac.uk/research/projects/poplog/examples

  "Thinky" programming is concerned with trying to get machines to perceive, reason, plan, develop motives and preferences, understand language, communicate in language, use diagrams for reasoning, make mathematical discoveries, develop theories, play various games, experience various feelings, and in some cases control physical behaviours on the basis of understanding the possibilities and constraints in a physical situation ... and more.

  The emphasis is not on simply producing observable behaviours, as in many programming exercises, but in producing analogues to mental processes in humans and other animals. Of course some of these mental processes can lead to behaviours, including use of language, playing games, and achievement of complex goals.

  AI researchers working on these topics have found it useful to develop programming languages with special support for the mechanisms required for "thinky" programming and the CAS-AI sub-group provides an introduction to such programming.

  This web site emphasises Programming in Pop11, which is part of the Poplog system, (http://en.wikipedia.org/wiki/POP-11), since that is the AI language best known to the maintainer, and it includes many teaching resources, at various levels, from absolute beginner programmers upwards.

  Other AI languages and AI teaching resources can be added (Please send links and accompanying text). Prolog and Common Lisp are included in the Poplog system mentioned below http://en.wikipedia.org/wiki/Poplog.

  The CAS-AI email list and the resources listed below are intended to serve the goals of introducing AI as science in schools, mainly by helping teachers gain some experience of AI programming. It is hoped that teachers who are interested can then go on to contribute to development of new versions of the teaching materials suited to their own pupils, using other programming languages if appropriate.

• The cas-ai mailing list
  Information about the CAS-AI mailing list and instructions for joining can be found at https://mailman.cs.bham.ac.uk/mailman/listinfo/cas-ai
  You can also subscribe by email as explained there. Send a message to cas-ai-request@cs.bham.ac.uk with Subject: subscribe
  Archived messages: http://mailman.cs.bham.ac.uk/archives/cas-ai/
  Non-members can read the message archives but cannot post to the list.

  Membership of the list is not restricted to CAS members: others who wish to join can follow instructions below (including submitting the form). However, joining CAS is highly recommended if you wish either to make a contribution to improving and broadening the quality of computing education, or to find out what others are doing about this.

• Form for list joiners
  If you decide to join, please look at this: Information form for new members.
  Please email the form to A.Sloman, as instructed.
  The form provides information about member expertise, resources, hoped for benefits, offers of help, school or institution, location....

---

Materials related to AI programming in Pop11 (an AI language that grows with the user)

• Materials for beginners (including teachers)
  • Some experimental Video tutorials (podcasts?) on AI concepts and techniques
    I have begun to produce some unscripted video tutorials illustrating AI programming concepts.
    They are uploaded to YouTube where they should be viewable in most browsers.

    A list of available tutorials, which will be reorganised when more videos are available is
    http://www.cs.bham.ac.uk/research/projects/poplog/cas-ai/video-tutorials.html
    or http://tinyurl.com/PopVidTut

    The originals, created using 'recordmydesktop' running on linux (Fedora 15/16), are in this directory
    http://www.cs.bham.ac.uk/research/projects/cogaff/tutorials
    The .ogv originals have better quality than the YouTube versions -- and should be playable using 'VLC' and some other media players.

  • Getting started with the Ved/XVed editor.
    • TEACH MINIVED (Minimal introduction)
    • TEACH QUICKVED (More detail on the editor)
    • TEACH ESSENTIALKEYS (Keyboard combinations for use without Function Keys)
    • TEACH BUFFERS (using several Ved buffers)

  • Some introductory AI-ish programming exercises (text only)
    This selection of text-based tutorials on Pop11 its editor and some aspects of AI is intended mainly for teachers new to AI languages.
    They can all be used on the text only version of poplog that runs on MSwindows, as well as being available on the remotely accessible Amazon based system described below.

    Some of the teaching examples that make use of the graphical facilities in Pop11, and require use of linux locally, or used remotely, or in a virtual linux, are listed below.

    • TEACH MISHAPS
    • TEACH SIMPLEMATHS Simple introduction to arithmetical operations
    • TEACH RIVER For absolute beginner programmers.
    • TEACH MATCHES Introduction to use of pattern matching in AI programming
      A speciality of Pop11. Presupposed by many of the subsequent files.
    • TEACH CHAT0 Getting started on a Simple chatbot. Playing with one.
    • TEACH CHAT1 The architecture of a Chatbot. Mark1 chatbot.
      More on the design of a simple chatbot, emphasising the architecture, to follow.
    • TEACH EVANS Introduction to designing and using a program that can reason about partial analogies between simple stories.
    • TEACH DATABASE Introduction to use of the Pop11 database, interrogated using pattern matching, as a simple tool for representing structured and changing knowledge.

    • Non-video teach files:
      • TEACH GRAM1
      • TEACH GRAM2

        .... to be continued ...

      • TEACH STORYGRAMMAR Using the Pop11 grammar library to generate and analyse simple stories and haikus.
    • Jocelyn Paine's Poplog page
      This web page by Jocelyn Paine in Oxford has more examples. I don't know if he is still teaching this material.
    • .... to be continued ...

  • Pop11 Syntax (an introductory subset)
    popcore.pdf (Printable version)
    popcore.txt
    Tiny introduction to a subset of features of the language pop11, for beginners.
    Plain text and pdf versions have the same contents.

    For a more detailed overview of Pop11 syntax see the Primer, below

  • Examples using graphics
    The previous examples could be used in a text only environment. The following make use of the 2-D graphical facilities in Pop11, which require use of linux poplog. This is available accessible remotely, or can be installed on a local linux system, or can be installed on windows or a mac using VirtualBox, as explained here.
    • TEACH GSTART
      Some tutorial examples using graphics. (To be revised and extended.)
    • TEACH FACES Making happy, sad, etc. faces out of coloured blobs.
    • TEACH RC_GRAPHIC
      Introduction to the pop11 2-D Relative-Coordinates (RC) graphical tools.
      Further illustrations are shown here.
    • Examples from TEACH RC_GRAPHPLOT (drawing graphs).
      (Developed by David Young at Sussex University.)
    • The RC_NEURAL demo
      Illustrating how to build and run a simple neural net constructor that shows graphically what it is doing.
      (By Riccardo Poli, now at Essex University.)
    • The SimAgent Toolkit
      This can be used by more advanced learners to produce simulations of interacting 2-D (more or less "thinky") agents.
      Some non-interactive movies showing student projects running are available here. The TEACH SIM_AGENT file included in Poplog helps learners get started.
    • ... To be continued ...

  • Web site with introductions to "thinky programming" using Pop11 (Graphical and non-graphical)

    Including a sample of a complete pop11 program, written as a project by an MSc Student (Neil Massey) in 1999, with many explanatory comments.
    This may be useful for CAS-AI members with a lot of programming experience.
    The program plays othello with a graphical interface:

    • Browsable pop-othello files
    • Overview file with suggestions for further work.
    • Download the othello files (ZIP)
    
    More examples are on the web site explaining how to get poplog working in VirtualBox.

  • Some experimental Video tutorials (podcasts?) on AI concepts and techniques
    illustrated using Pop11, have been placed on YouTube here. New ones will be added.
    The originals, and some draft new ones are in
    http://www.cs.bham.ac.uk/research/projects/cogaff/tutorials/
    The .ogv originals will have better quality than the YouTube versions.

  • A recorded 'Teach share' presentation
    on the need to teach programming as a way of increasing scientific understanding of natural information processing systems is here
    (with about 80 seconds of silence after the introduction because microphone was accidentally switched off):
    http://bit.ly/iVhp0i
    It includes demonstrations of some of the above teach files.
    The PDF slides supporting the presentation are in
    http://www.cs.bham.ac.uk/research/projects/cogaff/talks/#cas11
    

  • More on available teach files
    For more information use the TEACH INDEX command in the Editor (Ved)
    ENTER teach index RETURN.

  ---

• New version of POP11 PRIMER (Version 4)
  (The primer is written for people with prior programming experience.)
  Others should use introductory "teach" files, such as TEACH RIVER
  and materials listed below.)
  • pop11-primer.html
    Version4 of Pop11 Primer (HTML)
    (Still work in progress: links to be inserted.)
  • pop11-primer.pdf
    Printable Version 4 of Pop11 Primer (PDF) about 287 pages.
  • pop11-primer-2page.pdf
    Printable Version 4 of Pop11 Primer (PDF) two pages per sheet. about 144 pages.
  • pop11-primer.txt
    Plain text version of primer, for use in poplog editor Ved/XVed.
    Table of contents works with 'ENTER g' editor command.
    All the examples can be run in the editor - e.g. mark a line or a range and then compile it, then edit it and recompile,
    or copy and edit and compile.
    (This works for all the Poplog documentation included with with the system. The editor also knows how to find help, or show libraries.)
  • primer-version-3.txt
    Old version of pop11 primer, for comparison.
    The new version has some corrections and added sections.
  • HELP MATCHES Summary of Pop-11 matcher facilities.

  • For more information use the HELP INDEX command in the Editor
    ENTER help index RETURN.

• Proposal for teaching AI/Cognitive Science at A-Level AS-level
  A-Level and AS-Level AI/Cognitive Science proposal (HTML)
  A-Level and AS-Level AI/Cognitive Science proposal (PDF)

AI Toolkit and teaching materials on Amazon Cloud

• Members who join the CAS-AI list can apply for remote access to a Virtual computer running linux (Ubuntu) on the Amazon "cloud"), to gain experience of the kinds of AI programming mentioned above ("thinky programming").
  The virtual linux machine has been kindly made available by Dr. Lee Gillam at Surrey University.
  Anyone wishing to use it should fill in and submit the above form for new members.
  The linux machine includes a centrally managed powerful AI learning and teaching environment (Poplog/Pop11), on which users can gain experience. (It is hoped that this will enable teachers to get started on AI, and then develop their own teaching materials.)

  After gaining experience, some may wish to install the system locally either on a local linux system or using VirtualBox on a Windows PC or Mac, as described in
  http://www.cs.bham.ac.uk/research/projects/poplog/ova/

• Learning AI on a remote Linux machine.
  The main tools I know how to use for introducing AI use Linux and its graphical system (the X window System)
  One benefit is that the graphical applications can be run remotely, using a local X server on the user's machine.
  XMing is available for that purpose on machines running MSWindows.
  Converting for use in a browser would be technically difficult, but may be possible. Volunteers welcome. (The technology is already in XMing.)

  However, there are many programming teach files that do not require the graphical facilities and those can be read in a plain text window, e.g. an 'xterm' window if you are using linux and have logged in remotely, or in a PuTTY window on MSWindows, if you have used PutTTY to log in remotely. A list of example tutorials requiring only a plain text interface will be added shortly.
  You can download PuTTY or some alternatives, here.

  For people who don't wish to access the remote Amazon VM, or do not have adequate bandwidth for remote use, can use all the tools and teaching materials locally in linux or windows, the latter with the aid of VirtualBox.

  Although not strictly necessary at the beginning, people who wish to run all the examples and explore the programming environment will need to learn at least some basic unix commands. Information about the main features of Unix/Linux for novice users can be accessed below.

---

Useful links for people not familiar with linux/unix.

• Unix/Linux tutorial at Surrey university
• unix-intro.txt
  A brief introduction to unix and its history and ways of using it, equally applicable to linux. Will be updated.
  Introduces some general ideas about logging in, use of a window manager, file structure, and a few useful commands.
  Originally written for Birmingham students. Will be updated for cas-ai
• unix-commands.txt A useful subset of command-line (shell) commands for linux/unix users
• shell-commands.txt
  more detailed overview of commands, and some general information.
• unixhelp.html
  A collection of help files for unix, mostly relevant to linux, originally produced at Edinburgh university some time ago.

---

Other AI Resources (To be expanded)

• http://www.poplog.cs.rdg.ac.uk/
  The Reading University Poplog online documentation pages
  BEWARE: do not click on the "Poplog forum link", as www.popforum.org is now a dreadful advertising site that will attempt to take over your screen.
• http://www.aaai.org/aitopics/
  Web site provided by AAAI (Association for the advancement of AI).
• http://www.aisb.org.uk/
  UK-based AI society - founded 1969, the first AI society in the world. Organizes a conference every year. http://www.aisb.org.uk/convention
• http://www.bcs-sgai.org/
  BCS SIGAI - another UK-based AI society. Unfortunately ONLY members of BCS can join.
  (A totally daft policy, excluding many people who are potentially interested in AI but have no reason to join BCS! [Opinion of A.S.])
• http://blip.tv/pycon-us-videos-2009-2010-2011/easy-ai-with-python-1957038
  Survey several basic AI techniques implemented with short, open-source Python code recipes.

  Please suggest additions (to A.Sloman).

---

NOTE:
Later I'll add here a link to a short explanation of why I think it is very important to start teaching about architectures as well as algorithms and data-structures when teaching about programming. Without the concept of an architecture you cannot think clearly about the whole system you are attempting to design and what it is for.

---

INSTALLED: 17 Oct 2011
UPDATED: 19 Oct 2011; 30 Oct 2011; 5 Nov 2011; ... 16 Nov 2011; 22 Dec 2011; 9 Mar 2014;10 Jul 2015
Maintained by: Aaron Sloman
Email: a.sloman@cs.bham.ac.uk
http://www.cs.bham.ac.uk/research/projects/poplog/freepoplog.html
http://www.cs.bham.ac.uk/research/projects/poplog/packages/simagent.html

-

-

-

-

-

-

-

-

-

-

-

-

-

-