Research map / Graph learning / Knowledge and temporal graphs
LowFER: Low-rank Bilinear Pooling for Link Prediction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use LowFER-k* instead of TuckER | link prediction · FB15k | MRR: 0.795 -> 0.824 | Helped |
| Use LowFER-k* instead of TuckER | link prediction · FB15k-237 | MRR: 0.358 -> 0.359 | Helped |
| Use LowFER-k* instead of TuckER | link prediction · WN18 and WN18RR | MRR: WN18RR: 0.470 -> 0.465; WN18: 0.953 -> 0.950 | Hurt |
| Use LowFER-k* instead of HypER | link prediction · YAGO3-10 | MRR: 0.533 -> 0.537 | Helped |
| Increase factorization rank from 1 to 100 (ablation) | LowFER with d_e=d_r=200 · FB15k | MRR: 0.634 -> 0.779 | Helped |
| Increase factorization rank from 50 to 100 (ablation) | LowFER with d_e=d_r=200 · FB15k | MRR: 0.776 -> 0.779 | No clear effect |
| Add l2-regularization 0.0005 (ablation) | LowFER-k* with d_e=200, d_r=50, k=150 · FB15k | MRR: 0.795 -> 0.802 | Helped |
| Add tanh non-linearity (ablation) | LowFER-k* · FB15k-237 | MRR: 0.359 -> 0.345 | Hurt |
| Add tanh non-linearity (ablation) | LowFER-k* · FB15k | MRR: 0.824 -> 0.818 | Hurt |
| Add tanh non-linearity (ablation) | LowFER-k* · WN18RR | MRR: 0.465 -> 0.457 | Hurt |
| Add tanh non-linearity (ablation) | LowFER-k* · WN18 | MRR: 0.950 -> 0.950 | No clear effect |
| Remove power and l2 normalization before summation pooling (ablation) | LowFER · link prediction experiments | performance | No clear effect |
Closest papers by meaning
- On Multi-Relational Link Prediction With Bilinear Models (AAAI 2018 · similarity 0.91)
- SimplE Embedding for Link Prediction in Knowledge Graphs (NeurIPS 2018 · similarity 0.87)
- Tensor Decompositions for Temporal Knowledge Base Completion (ICLR 2020 · similarity 0.86)
- Relational Deep Learning: A Deep Latent Variable Model for Link Prediction (AAAI 2017 · similarity 0.86)
- Hierarchical Relational Learning for Few-Shot Knowledge Graph Completion (ICLR 2023 · similarity 0.85)
- Replacing Paths with Connection-Biased Attention for Knowledge Graph Completion (AAAI 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.