Submission Instructions
Benchmark Submissions
Main, Approximate, and ABA Tracks. To submit benchmark AFs or ABA frameworks, you will need to prepare an archive containing a set of AF or ABA frameworks in the corresponding input format, see Input format here. The set of benchmarks must be accompanied by a description of the benchmarks (preferrably in the format described under rules). Benchmarks of different levels of hardness and any number of benchmarks are of interest. (However, note that (i) the ranking schemes employed in the competition are based on the number of solved instances within a specific per-instance time-limit, and hence challenging benchmarks for currently available solvers may be of particular interest; and that (ii) the number on benchmarks used from a particular domain may be capped.) You may also contribute by submitting a benchmark generator software implementation, together with suggestions for suitable parameter values for generating interesting benchmark instances. Send a link to the archive by email to grp-iccma23@helsinki.fi using the subject title "ICCMA23 benchmark submission" by April 30, 2023 the latest.
Dynamic Track. To submit a benchmark application for the dynamic track, you need to prepare a Python program which employs the the Python version of IPAFAIR (ipafair.py), an API for incremental reasoning in abstract argumentation. In addition to adhering to the API, the program must fulfill the following requirements:
- The program takes as input a single command-line argument corresponding to a single input file.
- In case the input file is not an APX file containing an input AF, you must also provide a selection of input files. Note that an APX file may be directly passed to the AFSolver instance upon initialization.
- Note that the semantics must be specified when initializing the AFSolver instance. Allowed values are CO (complete), PR (preferred), and ST (stable).
- If the semantics is CO, only solve_cred() calls are allowed. If the semantics is PR, only solve_skept() calls are allowed. For ST, call either solve_cred() or solve_skept(), but do not combine both.
- The calls to solve_cred() or solve_skept() must contain a single query argument, which may however be changed between different calls.
Solver Submissions
To submit a solver you will need to prepare a zip or tar file (the suffix needs to be .zip, .tar, tar.gz, or .tgz) containing the solver source code, necessary third-party libraries, a README with detailed instructions on building the solver, and LaTeX sources and a compiled PDF of the solver description following the instructions and format described under under rules. Send the archive by email to grp-iccma23@helsinki.fi using the subject title "ICCMA23 solver submission" by April 30, 2023 the latest.
Dynamic Track. To submit a solver to the dynamic track, implement the functions defined in the the Python version of IPAFAIR (ipafair.py), an API for incremental reasoning in abstract argumentation, as described in the comments of the API. Note especially that if your solver does not provide support for a particular sequence of calls, the solver state must be changed to ERROR. See the examples (solver.py and solver.cpp) in the git repository for how to implement the Python API using a dynamic library written in C and the C version of IPAFAIR (ipafair.h). Please do not hesitate to contact the organizers by email in case you have any questions about your solver submission.