Research map / Vision-language and multimodal / Human and embodied modeling
GLoMo: Unsupervised Learning of Transferable Relational Graphs
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 |
|---|---|---|---|
| add unsupervised transferred graphs on top of GloVe features | SQuAD baseline · SQuAD GloVe | EM: 69.33 -> 70.84 | Helped |
| add transferred graphs on top of ELMo features | SQuAD ELMo baseline · SQuAD ELMo | EM: 74.75 -> 76.00 | Helped |
| add transferred graphs | MNLI baseline · MNLI matched | accuracy: 77.14 -> 78.32 | Helped |
| add transferred graphs to image model | ResNet-18 · CIFAR-10 | accuracy: 90.93 -> 91.55 | Helped |
| decouple graph and feature networks (ablation) | GLoMo · SQuAD GloVe | EM: 70.45 -> 70.84 | Helped |
| add ReLU sparsity to graphs (ablation) | GLoMo · SQuAD GloVe | EM: 70.13 -> 70.84 | Helped |
| use hierarchical multi-layer graphs (ablation) | GLoMo · SQuAD GloVe | EM: 69.92 -> 70.84 | Helped |
| use unit-level objective instead of sentence-level skip-thought (ablation) | GLoMo · SQuAD GloVe | EM: 69.23 -> 70.84 | Helped |
| predict context of length D instead of one unit (ablation) | GLoMo · SQuAD GloVe | EM: 69.92 -> 70.84 | Helped |
| use learned graphs instead of uniform graphs (ablation) | GLoMo · SQuAD GloVe | EM: 69.48 -> 70.84 | Helped |
| use learned graphs instead of uniform graphs (ablation) | GLoMo ResNet-18 · CIFAR-10 | accuracy: 91.07 -> 91.55 | Helped |
| remove sparsity on MNLI mismatched (ablation) | GLoMo · MNLI mismatched | accuracy: 78.00 -> 77.75 | Mixed |
Closest papers by meaning
- Transfer Learning for Deep Learning on Graph-Structured Data (AAAI 2017 · similarity 0.87)
- Learning Graphs for Knowledge Transfer With Limited Labels (CVPR 2021 · similarity 0.86)
- UniGTE: Unified Graph–Text Encoding for Zero-Shot Generalization across Graph Tasks and Domains (NeurIPS 2025 · similarity 0.85)
- GOFA: A Generative One-For-All Model for Joint Graph Language Modeling (ICLR 2025 · similarity 0.84)
- Contextual Structure Knowledge Transfer for Graph Neural Networks (AAAI 2025 · similarity 0.84)
- G-Adapter: Towards Structure-Aware Parameter-Efficient Transfer Learning for Graph Transformer Networks (AAAI 2024 · 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.