# Code

## Tractable Probabilistic Models and Deep Learning

**dCSN**: Learning Cut-Set Networks.**SPyN**: Routines for inference and learning with Sum-Product Networks in python. Code available on GitHub.

## Statistical Relational Learning

**LEMUR**: LEarning with a Monte carlo Upgrade of tRee search. Searches the space of possible theories using a Monte Carlo Tree Search (MCTS) algorithm. The algorithm is available in the git repository https://gitorious.org/yap-git/yap-lemur which is a user clone of the Yap 6.2.2 repository. To download it, please do*git clone git://gitorious.org/yap-git/yap-lemur.git*. Code and references avaible on https://sites.google.com/a/unife.it/ml/lemur.**RIB**: Relational Information Bottleneck. Applying the Information Bottleneck Approach to SRL: Learning LPAD Parameters. The algorithm RIB is distributed with YAP Prolog in the folder packages/cplint/rib. Code and references on https://sites.google.com/a/unife.it/rib/.**Lynx**: Probabilistic Models for Relational (Sequence). Learning. Code on https://github.com/nicoladimauro/lynx.

## Clustering-Grouping Problem

**ELK**: Algorithms for cooperative grouping (partitioning) problem. Code on https://github.com/nicoladimauro/elk.

## Inductive Logic Programming

- MDLS. An Inductive Logic Programming (ILP) algorithm for discovering first-order (DATALOG) maximal frequent patterns in multi-dimensional relational sequences.
- INTHELEX: INcremental THeory Learner from EXamples. An incremental learning system for the induction of hierarchical logic theories expressed as sets of Datalog clauses under Object Identity from positive and negative examples.
- A matching procedure for theta-subsumption. A procedure for the complete computation of all solutions to theta-subsumption between Horn clauses
- SPROL. Learner of relational concepts based on a stochastic propositionalization technique.
- RBK (Random Backtracking Depth First Search). A randomization strategy for ILP, similar to a restart strategy, able to randomly jump in the clause lattice to skip some portion of it.
- MILE (Mode-Declarations Incremental Learner from Examples). An incremental algorithm that induces mode declarations from instances of relational concept.