DRUM: End-To-End Differentiable Rule Mining On Knowledge Graphs
By A. Sadeghian, Mohammadreza Armandpour, Patrick Ding, Daisy Zhe Wang. Neural Information Processing Systems, 2019
414 citations recorded by the source corpus.
In this paper, we study the problem of learning probabilistic logical rules for inductive and interpretable link prediction. Despite the importance of inductive link prediction, most previous works focused on transductive link prediction and cannot manage previously unseen entities. Moreover, they are black-box models that are not easily explainable for humans. We propose DRUM, a scalable and differentiable approach for mining first-order logical rules from knowledge graphs that resolves these problems. We motivate our method by making a connection between learning confidence scores for each rule and low-rank tensor approximation. DRUM uses bidirectional RNNs to share useful information across the tasks of learning rules for different relations. We also empirically demonstrate the efficiency of DRUM over existing rule mining methods for inductive link prediction on a variety of benchmark datasets.
Subfields: Artificial Intelligence, Information Systems
View Paper: https://www.semanticscholar.org/paper/7445966ded8f38cac487314e8c6b71cdf327b047
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- William L. Hamilton
- Jian Tang
- Zhaocheng Zhu
- E. Denis
- Komal K. Teru
- Zuobai Zhang
- Louis-Pascal Xhonneux
- Yue Liu
In Information Systems
Citations in the corpus
- A Survey of Knowledge Graph Reasoning on Graph Types: Static, Dynamic, and Multi-Modal (2022, methodology): “Meanwhile, an end-to-end differentiable rule-based model is proposed in DRUM [144].”
- Knowledge Graphs (2020, neutral): “DRUM [116] also learns path-like rules, where, observing that some edge labels are more/less likely to follow others – for example, flight should not be followed by capital in the graph of Figure 15 as the join will be empty – the system uses bidirectional recurrent neural networks (a technique for ”
- Inductive Relation Prediction by Subgraph Reasoning (2019, methodology): “At a closer inspection, the previous differentiable methods (Neural-LP and DRUM) perform significantly worse than GraIL.”
- Neural Bellman-Ford Networks: A General Graph Neural Network Framework for Link Prediction (2021, methodology): “1 mins DRUM [46] SEAL [73] / X X O(|E|d(2)) ≈1 month GraIL [55]”