Inductive logic programming at 30: a new introduction
Paper • 2008.07912 • Published
The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
A collection of 55 relational learning and inductive logic programming (ILP) datasets.
| Dataset | Domain | Description | Tasks |
|---|---|---|---|
1d |
Visual Reasoning | 1D-ARC grid transformation tasks (Xu et al., AAAI 2023) | 18 types × 50 instances |
aids |
Molecular | Antiviral activity against HIV (Riesen & Bunke, SSPR/SPR 2008) | 1 |
alzheimer |
Molecular | Drug design for Alzheimer's disease, source attributed to King et al. 1995 | 4 |
arc1 |
Visual Reasoning | Abstraction and Reasoning Corpus (Chollet, arXiv 2019) | 2 |
arc2 |
Visual Reasoning | Abstraction and Reasoning Corpus (Chollet, arXiv 2019) | 1120 |
biodegradability |
Molecular | Aerobic aqueous biodegradability prediction (Džeroski et al., Applied AI 1998) | 1 |
bongard |
Visual Reasoning | Visual concept learning in the Bongard style (Nie et al., NeurIPS 2020) | 12 |
bzr |
Molecular | Benzodiazepine receptor binding (Sutherland et al., J. Chem. Inf. 2003) | 1 |
chebi_1624 |
Molecular | ChEBI ontology molecular classification (Hastings et al., NAR 2016) | 1808 |
chess |
Games | King-Rook vs King endgame (Michie et al., UCI ML Repository 1994) | 3 |
chess_tactics |
Games | Tactical theme classification from the Lichess open puzzle database, CC0 | 9 |
clevr |
Visual Reasoning | Visual reasoning tasks derived from CLEVR (Johnson et al., CVPR 2017) | 130 |
conceptarc |
Visual Reasoning | ConceptARC benchmark (Moskvichev et al., arXiv 2023) | 16 |
conceptnet |
Knowledge Graph | ConceptNet 5.5 commonsense relation prediction (Speer et al., AAAI 2017) | 10 |
cora |
Graph Structure | Author-field token classification in bibliographic references (McCallum et al., 2000; Kok & Domingos, ICML 2005) | 1 |
cox2 |
Molecular | Cyclooxygenase-2 binding (Sutherland et al., J. Chem. Inf. 2003) | 1 |
dhfr |
Molecular | Dihydrofolate reductase binding (Sutherland et al., J. Chem. Inf. 2003) | 1 |
drugdrug |
Graph Structure | Drug-drug interaction prediction (Dhami et al., Smart Health 2018) | 1 |
ecoli |
Molecular | Structure-activity relationships: E. coli, pyrimidines, triazines (Muggleton et al., ML Workshop 1992) | 3 |
fb15k |
Knowledge Graph | Freebase 15K link prediction (Bordes et al., NeurIPS 2013) | 237 |
flashfill |
Program Synthesis | FlashFill Excel string transformations (Gulwani, POPL 2011) | 15 |
geoi_lp |
Other / Synthetic | GeoILP synthetic rule induction benchmark (Chen et al., ICLR 2025) | 1042 |
grapheval |
Graph Structure | LeetCode graph algorithm problems, GraphEval36K benchmark (Wu et al., arXiv 2024) | 40 |
hepatitis |
Molecular | Hepatitis subtype classification (Motl & Schulte, arXiv 2015) | 1 |
iggp |
Games | Inductive General Game Playing — 50 diverse games (Cropper et al., Machine Learning 2020) | 340 |
imdb |
Knowledge Graph | IMDB movie industry relational learning (Blockeel & De Raedt, AI 1998) | 3 |
iparc |
Visual Reasoning | ILP-flavoured ARC challenge (Challa et al., ILP 2022) | 6 |
jr |
Program Synthesis | List function tasks (Rule et al., Nature Communications 2024) | 250 |
kandinsky |
Visual Reasoning | Kandinsky pattern concept learning (Müller & Holzinger, AIJ 2021) | 24 |
kinship |
Knowledge Graph | Family relationship learning (Hinton, CogSci 1986) | 7 |
legal2logic |
Other / Synthetic | Legal2Proleg statutory rule learning (Xue et al., arXiv 2026) | 18 |
lpp |
Games | Logic program policies for games (Silver et al., AAAI 2020) | 5 |
mesh |
Graph Structure | Mesh dataset from the IC London ILP dataset collection | 1 |
mini-arc |
Visual Reasoning | Miniature ARC tasks (Kim et al., NeurIPS Workshop 2022) | 149 |
mol |
Molecular | Molecular property prediction, OGB collection (Hu et al., NeurIPS 2020) | 61 |
mutag |
Molecular | Mutagenicity of nitroaromatic compounds (Debnath et al., J. Med. Chem. 1991) | 1 |
nations |
Knowledge Graph | Political/economic/military relations between 14 countries (Rummel, Sage 1972; Kok & Domingos, ICML 2007) | 55 |
nci1 |
Molecular | NCI anti-cancer activity benchmark (Wale et al., KAIS 2008) | 1 |
nci109 |
Molecular | NCI anti-cancer activity benchmark, companion to NCI1 (Wale et al., KAIS 2008) | 1 |
peptides_func |
Molecular | Peptide functional classification, LRGB (Dwivedi et al., NeurIPS 2022) | 10 |
protein |
Molecular | Protein function classification from the IC London ILP dataset collection | 1 |
PTC |
Molecular | Predictive Toxicology Challenge carcinogenicity (Helma et al., Bioinformatics 2001) | 1 |
qm9 |
Molecular | Quantum chemistry molecular properties (Ramakrishnan et al., Sci. Data 2014) | 15 |
relbench |
Other / Synthetic | RelBench relational database prediction tasks (Robinson et al., arXiv 2024) | 59 |
satellite |
Graph Structure | Satellite dataset from the IC London ILP dataset collection | 1 |
stepgame |
Spatial Reasoning | Multi-step spatial relational reasoning (Shi et al., AAAI 2022) | 9 |
strings |
Program Synthesis | Character-level string transformations (Cropper, IJCAI 2019) | 329 |
tox21 |
Molecular | Toxicology across 12 assays, OGB (Hu et al., NeurIPS 2020) | 12 |
toxicology |
Molecular | Carcinogenicity in rodents, CTU Prague (Helma et al., Bioinformatics 2001) | 1 |
trains |
Games | Trains classification (Larson & Michalski, SIGART 1977) | 4 |
uwcse |
Knowledge Graph | UW-CSE university department advisedby prediction (Richardson & Domingos, ML 2006) |
1 |
wn18rr |
Knowledge Graph | WordNet 18 Reduced Relations link prediction (Dettmers et al., AAAI 2018) | 11 |
yago3_10 |
Knowledge Graph | YAGO3 knowledge graph, 10 high-frequency relations (Mahdisoltani et al., CIDR 2015) | 10 |
yeast |
Molecular | Yeast gene function classification (Struyf et al., ECML 2006) | 1 |
zendo |
Visual Reasoning | Zendo concept learning game (Bramley et al., CogSci 2018) | 14 |
zinc250k |
Molecular | ZINC250k drug-likeness properties (Gómez-Bombarelli et al., ACS Cent. Sci. 2018) | 3 |
Total: 56 datasets, 6131+ tasks
Each task directory contains three files:
bk.pl — background knowledge (ground facts)exs.pl — positive and negative examplesbias.pl — hypothesis search space (head/body predicates, types, formatted for the Popper system.)