Luc Steels: Can Machines Be Creative Enough To Invent Their Own Language?

Speaker: Luc Steels, CSIC-UPF Barcelona and director Sony CS Laboratory Paris.
Title: Can Machines Be Creative Enough To Invent Their Own Language?
Time: Monday 5th of November at 18:00
Place: Egget, Student Centre
Followed by debate with Terje Lohndal, “the Mozart of linguistics”.

For the past decade, I have been doing experiments with my team trying to see in how far we could program a group of robots to autonomously create and learn from each other a symbolic communication system with similar properties as we find in human languages. We have managed to do some breakthrough experiments showing the emergence of lexicons for colour, space, and actions with robots playing language games and have made significant headway on the question of the emergence of grammar. This talk shows through some stunning video clips what we have been able to achieve and discusses what this tells us about the nature of our own intelligence, the mechanisms underlying cultural evolution and the future of artificial intelligence.

Luc Steels is an ICREA research professor at the Institut de Biologia Evolutiva (CSIC-UPF) in Barcelona and director of the Sony Computer Science Laboratory in Paris.

Terje Lohndal is an associate professor in linguistics at NTNU.

Marcus du Sautoy, Simonyi professor for the public understanding of science and a professor of mathematics at the University of Oxford has the following to say about our next speaker, Luc Steels, in the Guardian feature story AI robot: how machine intelligence is evolving:

For me one of the most striking experiments in AI is the brainchild of the director of the Sony lab in Paris, Luc Steels. He has created machines that can evolve their own language. A population of 20 robots are first placed one by one in front of a mirror and they begin to explore the shapes they can make using their bodies in the mirror. Each time they make a shape they create a new word to denote the shape. For example the robot might choose to name the action of putting the left arm in a horizontal position. Each robot creates its own unique language for its own actions.

Thore Husfeldt: Let us calculate! — From Leibniz to Turing

This Friday, 22nd of June at 18.30, the first Alan Turing 2012 meeting at UiB will take place. At 18.30, we will show a pre-release version of the movie Codebreaker. At 19.30, there will be a talk by Thore Husfeldt, called “Let us calculate! — From Leibniz to Turing”. It will all take place at Kvarteret, Det Akademiske Kvarter, in the hall Tivoli.

CODEBREAKER tells the remarkable and tragic story of one of the 20th century’s most important people. Alan Turing set in motion the computer age and his World War II codebreaking helped turn the tide of the Second World War.

Instead of receiving accolades, Turing faced terrible persecution. In 1952, the British Government forced him to undergo chemical castration as punishment for his homosexuality. In despair, Turing committed suicide. He was only 41 years old.

Documentary elements seamlessly interconnect with drama scenes in CODEBREAKER to offer a three dimensional picture of Turing, his accomplishments, his tragic end, and his lasting legacy.

For a short trailer, visit Alan Turing Film Trailer.


Let us calculate! — From Leibniz to Turing

Thore Husfeldt, Lund University and ITU Copenhagen

I describe the intellectual history of computational thinking, beginning with the vision of Leibniz, an early advocate of rationalism, to solve differences of opinion by symbolic reasoning in a sufficiently strong formal system. From there, the road goes via Boole, Frege, Russel, and Gödel to heartbreak and catastrophe, and via Alan Turing to the triumph of the mechanical, algorithmic worldview of today. I end with an appraisal of the possible consequences of the algorithmic world for sex, lies, and videotapes, and the future of mankind.

Ken Regan: Scoping the Mind with Turing’s Chess Machine

Speaker: Kenneth W. Regan, University at Buffalo (SUNY)
Title: Scoping the Mind with Turing’s Chess Machine
Time: Tuesday 11th of September, 18:00.
Place: Teglverket, Kvarteret. Hosted by Upop, Studentersamfunnet.

Alan Turing’s dream of a champion chess machine aimed to reveal computational intelligence by simulating a signature activity of the human mind. Yet when losses by Garry Kasparov and other human champions were ascribed more to processors and speed than simulated thought, chess no longer seemed the “Drosophila of AI”. We regain Turing’s objectives by turning the machine onto its inventor, using computer analysis of human games to reveal patterns of human minds. Statistical analysis of tens of millions of pages of data argues these as mental laws:

  1. Humans judge moves not by difference in raw value as computers do, but in proportion to the
    overall advantage in the position.
  2. Skill as measured by Elo rating is simply linearly related to average error per move—depth
    of strategy and planning may be secondary.
  3. Procrastination and future-blind tendencies causing error are the same scross skill levels.
  4. Human skill has improved in similar manner to Olympic sports, contrary to widespread belief of “inflation” in chess ratings.
  5. Variation in performance obeys statistical laws at move level not just game level.
  6. Computers can catch humans cheating with (other) computers.

The talk will include lessons from work on cheating cases since 2006, issues in objective
skill assessment, and potential applications beyond chess. Some work is joint with Guy Haworth
and Giuseppe DiFatta (Univ. or Reading, UK) and Grandmaster Bartlomiej Macieja (Warsaw, Poland).

On the other side of the table Simen Agdestein is sitting, our own international grand master and football player.