Software

 

Machine learning

Some software is also available at https://github.com/ML-KULeuven.

Neurosymbolic AI and Probabilistic Programming:

  • DeepStochLog: is a neuro-symbolic framework that combines grammars, logic, probabilities and neural networks.
  • DeepProbLog: A programming language that seamlessly combines probabilistic logic programming and deep learning.
  • ProbLog is a probabilistic Prolog, a probabilistic logic programming language.
  • PySDD: Python wrapper for Sentential Decision Diagrams
  • Congrads: a Python toolbox that brings constraint-guided gradient descent capabilities to your machine learning projects
  • Klay: a Python library for evaluating sparse circuits on the GPU
  • NeSy: a differentiable shielding technique that can be used at training time to integrate user-defined constraints into the training process

Anomaly Detection:

  • Anomatools: Toolbox for anomaly detection
  • DTAIanomaly: A simple-to-use Python package for the development and analysis of time series anomaly detection techniques.

Prediction and Clustering:

  • COBRAS: Interactive clustering with pairwise constraints
  • DTAIDistance: Dynamic Time Warping and Time Series Clustering (Fast DTW implementation)
  • MERCS: Multi-Directional Ensemble of Regression and Classification treeS
  • SAR: The Selected At Random (SAR) assumption for learning from Positive and Unlabeled data
  • TIcE: estimating the label frequency in Positive and Unlabeled (PU) data, using tree induction.

Knowledge Representation

Knowledge Representation and Reasoning

  • IDP-Z3: A new system building on the ideas of IDP3, relying on the Z3 solver.
  • IDP: IDP allows models of systems and problems to be written in a specification language, which is a fragment of FO(·), and can perform several inference tasks such as model generation or querying on this specification.
  • MinisatID: The solver underlying IDP.
  • IDPDraw: IDPDraw is a tool for visualizing finite structures. It can be used to visualize the output of an ASP solver.
  • ConfigID: ConfigID is an API for the Java programming language enabling a Java application to use the IDP knowledge base system for configuration problems.
  • Asystem: An Abductive Constraint System
  • Interactive consultant: An interactive user interface on top of IDP-Z3.
  • FOLASP: Model expansion engine for FO(·) with back-end ASP solver Clingo
  • DMN-IDP: Integration of the tabular Decision Model and Notation standard with the IDP-Z3 solver, including a verification tool.
  • cDMN: extension of the DMN standard with constraints and an IDP-Z3 solver
  • pDMN: extension of the DMN standard with probabilistic reasoning to better express uncertainty in decision models.
  • (Typed) Feature Model - IDP: combines an intuitive feature modeling editor with IDP-Z3's Interactive Consultant interface.
  • FOLL-E: a tool to learn logical reasoning without the formal notation.

Constraint solving

  • CPMPy: a Constraint Programming and Modeling library in Python, based on numpy, with direct solver access.
  • Pindakaas: a library to transform pseudo-Boolean and integer constraints into conjunctive normal form.

Knowledge Graphs

  • MLSea: create and explore metadata about ML concepts
  • SCOOP: to create SHACL shapes from the data's schema, ontology and mappings
  • TorchicTab: automatically understands tables and provides annotations for these tables.

Sports Analytics

Old systems

  • 3DNK (3D Neighbourhood Kernel): a kernel function that can deal with 3D structures.
  • ACE (A Combined Engine): is the relational datamining system that implements a.o. the algorithms Tilde, Warmr, ICL, and RRL.
  • AMIE (Automatic Monitoring of Indoor Exercises): system for analysing and classifying physical exercises recorded with a Microsoft Kinect
  • BiQL a system for analyzing information networks.
  • Subtle/CellPhinder: A system to run Subgroup Discovery on time-lapse microscopy data in microbiology.
  • Claudien: a clausal discovery engine
  • Clus: a decision tree and rule induction system that implements the predictive clustering framework.
  • COBRA: a fast and simple method for active clustering with pairwise constraints
  • COBS: a system for constraint-based (or semi-supervised) clustering
  • COP-Kmeans: A Python implementation of COP-KMEANS algorithm
  • CP4IM: a system for mining frequent itemsets using constraint programming.
  • CURLED: an unsupervised representation learning approach for relational data.
  • C-FARMR: a system for mining free clauses.
  • Distributional Clauses (DC): Probabilistic logic language for inference, planning and learning in static and dynamic domains
  • DL8(.tar.gz): a constraint-based optimal decision tree learner.
  • DMax Chemistry Assistant a QSAR data mining system.
  • DucQ: A Python implementation of the QuAcq algorithm for constraint acquistion.
  • Experiment database: a database designed to store learning experiments in full detail, aimed at providing a convenient platform for the study of learning algorithms.
  • FOG: a system for mining outerplanar graph patterns under BBP subgraph isomorphism.
  • GC-FOVE: a system for performing lifted probabilistic inference by variable elimination.
  • GSSL: a randomized feature generation approach for Markov network structure learning.
  • HAL-ProbLog: a probabilistic logic programming language that extends the semantics and syntax of ProbLog to the continuous domain.
  • JSDD: Java wrapper for Sentential Decision Diagrams
  • kLogNLP: a kLog module for Graph Kernel-based Relational Learning of Natural Language.
  • LearnSDD: a Markov network structure learning that only learns tractable networks by using SDDs.
  • LLM: a structure learner for relational dependency networks in hybrid domains
  • MERCI: Identifying Discriminative Classification Based Motifs in Biological Sequences.
  • MIPS: A C++ Library for Graph-based Learning and Mining Algorithms.
  • NSPDK: Neighborhood Subgraphs Pairwise Distance Kernel.
  • onto2problog: a tool for ontology-mediated querying over probabilistic data for ontologies formulated in (essentially) OWL 2 EL
  • PIUS: Peptide Identification by Unbiased Search
  • ProbFoil: Probabilistic Rule Learner
  • PMCSFG: Pair-wise Maximum Common Subgraph Feature Generation.
  • PROFILE (Probabilistic First-Order Learning): a set of software tools for Statistical Relational Learning and Probabilistic ILP.
  • psipy: a python wrapper for the symbolic probabilistic inference engine underlying the PSI-Solver.
  • PyDC: a python wrapper that allows you to use Dynamic Distributional Clauses seamlessly in python.
  • pywmi: a framewoek and toolbox for probabilistic inference using weighted model integration
  • ReCeNT: a versatile framework for clustering relational data
  • Subtle YAP/SWI Prolog module for efficiently performing theta-subsumption tests and computing the least general generalization under theta-subsumption of two clauses
  • Taxify
  • TODTLER: a deep transfer learning algorithm for Markov logic networks.
  • WFOMC: a system for performing lifted probabilistic inference by first-order knowledge compilation.