ACL1 Version 1
ACL version 1 is a system that allows to learn
the program part of an abductive logic program and that produces the necessary
file for ICL so that relevant constraints can subsequently be learned.
Prerequisites:
The following software must be installed in your system:
Installation instructions:
-
download the ACL1 code: acl1_1.tar.gz
-
create a directory for your code and data files
-
copy acl1_1.tar.gz into it and decompress it
gunzip acl1_1.tar.gz
tar xvf acl1_1.tar
-
set an environmental variable for the execution of ICL
setenv ICL_APPL_CONF ''
Running the system:
-
prepare your data: you need two input files: (see input
file format)
-
<file_stem>.bg: the input file with the training set, background
knowledge and bias for the first phase of ACL1
-
<file_stem>.l: a file containing the definition of the bias
for learning constraints in the DLAB formalism used by ICL
-
load Sicstus
sicstus
-
load the acl1 code
| ?-[acl1_1].
-
start the induction with the command
the induction writes the output files <file_stem>.rules containing
the rules learned in the first phase and the file <file_stem>.kb
containing the input interpretations for ICL.
ICL needs an extra input file called <file_stem>.s
that contains the settings under which ICL is run. A set of standard
settings is provided in the file prototype.s that is automatically
copied by the acl1 code into <file_stem>.s. If
the standard setting do not give the expected result, they can be manually
changed by editing the <file_stem>.s file.
-
run ICL by calling
-
start the ICL learning process by giving the ICL command induce
(abbreviated i)
-
write the results with the command write_theory (abbreviate wt)
that
writes the learned constraints in the file <file_stem>.theory
Parameters setting:
A number of parameters can be set, both for ACL1 and for ICL. The ACL1
parameters are defined by a number of prolog facts and are described below.
The parmeter settings for ICL are included in the file <file_stem>.s
and are initialized with values useful for learning constraints. They can
be changed by referring to the ICL
manual for a description of the available settings.
The settings available in ACL1 are
-
verbosity(V): specifies the verbosity level. Set by default
to 3. According to the level of verbosity, the following information
is printed:
0: the output rules
1: every rule that is added to the theory
2: every rule added to the agenda
3: the agenda at each specializing step
-
beamsize(Beamsize): specifies the size of the beam. Set by default
to 5
-
nmax(Nmax): maximum number of specialization steps. Set
by default to 10
-
der_depth(Depth): depth of derivations. Set by default to
20
-
kt(K): threshold on K+ and K-. Set by default to 0.1
-
min_cov(MC): minimum number of examples that each clause must
cover. Set by default to 1
They can be set either by including a fact for them in the input file,
by changing the acl1_1.pl file and reloading it or by retract
and asserts.
For reporting impressions on the use of the system, bugs or new applications,
send an e-mail to Fabrizio Riguzzi friguzzi@deis.unibo.it.
Back to the main ACL page.