Search Results for author: Sebastijan Dumancic

Found 11 papers, 2 papers with code

DeepSaDe: Learning Neural Networks that Guarantee Domain Constraint Satisfaction

no code implementations2 Mar 2023 Kshitij Goyal, Sebastijan Dumancic, Hendrik Blockeel

In this paper, we present an approach to train neural networks which can enforce a wide variety of constraints and guarantee that the constraint is satisfied by all possible predictions.

SaDe: Learning Models that Provably Satisfy Domain Constraints

no code implementations1 Dec 2021 Kshitij Goyal, Sebastijan Dumancic, Hendrik Blockeel

In many real world applications of machine learning, models have to meet certain domain-based requirements that can be expressed as constraints (e. g., safety-critical constraints in autonomous driving systems).

Autonomous Driving

Feature Interactions in XGBoost

no code implementations11 Jul 2020 Kshitij Goyal, Sebastijan Dumancic, Hendrik Blockeel

Further, the improvement in the model structure can also lead to better interpretability.

Knowledge Refactoring for Inductive Program Synthesis

1 code implementation21 Apr 2020 Sebastijan Dumancic, Tias Guns, Andrew Cropper

We introduce the \textit{knowledge refactoring} problem, where the goal is to restructure a learner's knowledge base to reduce its size and to minimise redundancy in it.

Inductive logic programming Program induction +1

Learning Relational Representations with Auto-encoding Logic Programs

no code implementations29 Mar 2019 Sebastijan Dumancic, Tias Guns, Wannes Meert, Hendrik Blockeel

This framework, inspired by the auto-encoding principle, uses first-order logic as a data representation language, and the mapping between the original and latent representation is done by means of logic programs instead of neural networks.

Relational Reasoning Representation Learning

COBRA: A Fast and Simple Method for Active Clustering with Pairwise Constraints

no code implementations30 Jan 2018 Toon Van Craenendonck, Sebastijan Dumancic, Hendrik Blockeel

Clustering is inherently ill-posed: there often exist multiple valid clusterings of a single dataset, and without any additional information a clustering system has no way of knowing which clustering it should produce.

Clustering-Based Relational Unsupervised Representation Learning with an Explicit Distributed Representation

no code implementations28 Jun 2016 Sebastijan Dumancic, Hendrik Blockeel

The goal of unsupervised representation learning is to extract a new representation of data, such that solving many different tasks becomes easier.

General Classification Relational Reasoning +1

An expressive dissimilarity measure for relational clustering using neighbourhood trees

no code implementations29 Apr 2016 Sebastijan Dumancic, Hendrik Blockeel

It is the first measure to incorporate a wide variety of types of similarity, including similarity of attributes, similarity of relational context, and proximity in a hypergraph.

Cannot find the paper you are looking for? You can Submit a new open access paper.