Track A: AISC (Co-Chairs: John A. Campbell, Jacques Carette)

Symbolic computation can be roughly described as the study of algorithms which operate on expression trees. Another way to phrase this is to say that the denotational semantics of expressions trees is not fixed, but is rather context dependent. Expression simplification is probably the archetypal symbolic computation. Mathematically oriented software (such as the so-called computer algebra systems) have been doing this for decades, but not long thereafter, systems doing proof planning and theorem discovery also started doing the same; some attempts at knowledge management and 'expert systems' were also symbolic, but less successfully so. More recently, many different kinds of program analyses have gotten `symbolic', as well as some of the automated theorem proving (SMT, CAV, etc).

But a large number of the underlying problems solved by symbolic techniques are well known to be undecidable (never mind the many that are EXP-time complete, etc). Artificial Intelligence has been attacking many of these different sub-problems for quite some time, and has also built up a solid body of knowledge. In fact, most symbolic computation systems grew out of AI projects.

These two fields definitely intersect. One could say that in the intersection lies all those problems for which we have no decision procedures. In other words, decision procedures mark a definite phase shift in our understanding, but are not always possible. Yet we still want to solve certain problems, and must find 'other' means of (partial) solution. This is the fertile land which comprises the core of AISC.

Rather than try to exhaustively list topics of interest, it is simplest to say that AISC seeks work which advances the understanding of

1. Karlsruhe (Germany)
2. Cambridge (United Kingdom)
3. Steyr (Austria)
4. Plattsburgh (USA)
5. Madrid (Spain)
6. Marseille (France)
7. Hagenberg (Austria)
8. Beijing (China)
9. Birmingham (UK)
10. Paris (France)