Details
| Original language | English |
|---|---|
| Title of host publication | Machine Learning and Knowledge Discovery in Databases - International Workshops of ECML PKDD 2019 |
| Editors | Peggy Cellier, Kurt Driessens |
| Pages | 229-240 |
| Number of pages | 12 |
| Volume | 1 |
| Publication status | Published - 2020 |
| Event | 19th Joint European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2019 - Wurzburg, Germany Duration: 16 Sept 2019 → 20 Sept 2019 |
Publication series
| Name | Communications in Computer and Information Science |
|---|---|
| Volume | 1167 |
| ISSN (Print) | 1865-0929 |
| ISSN (electronic) | 1865-0937 |
Abstract
In this paper we propose and study the novel problem of explaining node embeddings by finding embedded human interpretable subspaces in already trained unsupervised node representation embeddings. We use an external knowledge base that is organized as a taxonomy of human-understandable concepts over entities as a guide to identify subspaces in node embeddings learned from an entity graph derived from Wikipedia. We propose a method that given a concept finds a linear transformation to a subspace where the structure of the concept is retained. Our initial experiments show that we obtain low error in finding fine-grained concepts.
Keywords
- Conceptual spaces, Interpretability, Node embeddings
ASJC Scopus subject areas
- Computer Science(all)
- General Computer Science
- Mathematics(all)
- General Mathematics
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
Machine Learning and Knowledge Discovery in Databases - International Workshops of ECML PKDD 2019. ed. / Peggy Cellier; Kurt Driessens. Vol. 1 2020. p. 229-240 (Communications in Computer and Information Science; Vol. 1167).
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research
}
TY - GEN
T1 - Finding Interpretable Concept Spaces in Node Embeddings Using Knowledge Bases
AU - Idahl, Maximilian
AU - Khosla, Megha
AU - Anand, Avishek
PY - 2020
Y1 - 2020
N2 - In this paper we propose and study the novel problem of explaining node embeddings by finding embedded human interpretable subspaces in already trained unsupervised node representation embeddings. We use an external knowledge base that is organized as a taxonomy of human-understandable concepts over entities as a guide to identify subspaces in node embeddings learned from an entity graph derived from Wikipedia. We propose a method that given a concept finds a linear transformation to a subspace where the structure of the concept is retained. Our initial experiments show that we obtain low error in finding fine-grained concepts.
AB - In this paper we propose and study the novel problem of explaining node embeddings by finding embedded human interpretable subspaces in already trained unsupervised node representation embeddings. We use an external knowledge base that is organized as a taxonomy of human-understandable concepts over entities as a guide to identify subspaces in node embeddings learned from an entity graph derived from Wikipedia. We propose a method that given a concept finds a linear transformation to a subspace where the structure of the concept is retained. Our initial experiments show that we obtain low error in finding fine-grained concepts.
KW - Conceptual spaces
KW - Interpretability
KW - Node embeddings
UR - http://www.scopus.com/inward/record.url?scp=85083711218&partnerID=8YFLogxK
U2 - 10.48550/arXiv.1910.05030
DO - 10.48550/arXiv.1910.05030
M3 - Conference contribution
AN - SCOPUS:85083711218
SN - 9783030438227
VL - 1
T3 - Communications in Computer and Information Science
SP - 229
EP - 240
BT - Machine Learning and Knowledge Discovery in Databases - International Workshops of ECML PKDD 2019
A2 - Cellier, Peggy
A2 - Driessens, Kurt
T2 - 19th Joint European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2019
Y2 - 16 September 2019 through 20 September 2019
ER -