AAAI-11, the Twenty-Fifth Conference on Artificial Intelligence

What the conference was

AAAI-11 was the twenty-fifth edition of the Conference on Artificial Intelligence, held at the Hyatt Regency San Francisco from August 7 to 11, 2011 ^[1]. Like the editions before it, it was a general meeting for the field rather than a single-topic workshop: the stated purpose was to promote research in artificial intelligence and to encourage scientific exchange among AI researchers, practitioners, scientists and engineers in related disciplines ^[1].

A conference of this kind works through a mix of formats. Papers are presented and discussed, senior figures give keynote addresses that survey where the field stands, and invited speakers bring in work from neighbouring areas. Much of the value lies outside the formal programme, in the conversations between people who would otherwise be working in separate labs and companies.

Who spoke

The keynote speaker was David Ferrucci of the IBM T. J. Watson Research Center ^[1]. A keynote is the slot in which a conference puts work it considers to have the broadest interest for the field, and it sets a theme that the rest of the week tends to circle around.

The 2011 Robert S. Engelmore Memorial Award Lecture was given by Ramon Lopez de Mantaras of the Artificial Intelligence Research Institute and the Spanish National Research Council ^[1]. The Engelmore lecture is a named address held in memory of a figure in the field, which makes it one of the conference's recurring landmarks.

Beyond the keynote and the named lecture, AAAI lists five invited speakers: Heinrich H. Buelthoff, Karrie Karahalios, Michael Kearns, Kurt Konolige and David Gunning ^[1]. The list spans quite different research backgrounds, which fits the general scope of the meeting.

What was left behind

Selected lectures and programmes from the conference were posted at VideoLectures.Net ^[1]. For a conference, this matters more than it might seem. Talks are normally visible only to the people in the room, and the audience is limited by travel and cost. Putting even a selection online turns a one-week event into material that students and researchers elsewhere can watch later, and it gives the field a record of how its central questions were being framed at that moment.

Why it matters

AAAI-11 sits in a long-running series, and the twenty-fifth edition is a marker of how established the annual general conference had become for artificial intelligence. Its programme shows the field's breadth at that point: a keynote from an industrial research laboratory, an award lecture from a public research institute, and invited talks from a mixed group of academic and industrial speakers ^[1]. For readers looking at the period afterwards, the conference is useful precisely as a snapshot — the topics that a broad AI audience was expected to care about, recorded in one place and partly preserved on video ^[1].