Research map / Learning theory and algorithms / Kernels and regression theory
Efficient Online Learning for Mapping Kernels on Linguistic Structures
Giovanni Da San Martino, Alessandro Sperduti, Fabio Aiolli, Alessandro Moschitti
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Represent the perceptron model as a DAG of shared subtrees (voted DAG perceptron) instead of the standard perceptron | Perceptron with SST+Poly3 kernel · PropBank boundary detection, 1M examples | training time (s): 22, 892.62 -> 15, 814.09 | Improved |
| Use DAG-based model representation instead of the standard perceptron model | Perceptron with SST+Poly3 kernel · PropBank boundary detection, 1M examples | nodes stored in model: 391, 240 -> 124, 333 | Improved |
| Increase training examples from 1 million to 2 million | Voted DAG Perceptron with SST kernel · PropBank boundary detection test | F1: 0.805 -> 0.814 | Improved |
| Train on the full training set instead of 2 million examples | Voted DAG Perceptron with SST kernel · PropBank boundary detection test | F1: 0.814 -> 0.822 | Improved |
| Combine the SST and degree 3 polynomial kernels | Voted DAG Perceptron · PropBank boundary detection test | F1: 0.822 -> 0.831 | Improved |
| Use the polynomial kernel or kernel combination instead of SST alone | Voted DAG Perceptron · PropBank boundary detection, full training set | training time (s): 51, 235.9 -> 178, 826.19 | Hurt |
| Voted DAG perceptron trained on 2M examples instead of SVM on 1M examples | SST kernel learner · PropBank boundary detection | training time: 7.5 days -> 4.3 hours | Improved |
Closest papers by meaning
- Convolutional Kernel Networks for Graph-Structured Data (ICML 2020 · similarity 0.85)
- Scalable Learning in Reproducing Kernel Krein Spaces (ICML 2019 · similarity 0.85)
- Discriminative Embeddings of Latent Variable Models for Structured Data (ICML 2016 · similarity 0.85)
- Graph-Aided Online Multi-Kernel Learning (JMLR 2023 · similarity 0.85)
- On the Target-kernel Alignment: a Unified Analysis with Kernel Complexity (NeurIPS 2024 · similarity 0.84)
- Frame Semantic Role Labeling Using Arbitrary-Order Conditional Random Fields (AAAI 2024 · similarity 0.83)
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.
