Name
Abstract | ||
|---|---|---|
Structured prediction provides a flexible modeling framework to deal with several relevant problems. Sequences, Trees, Disjoint Intervals and Matching are some useful examples of the type of structures we would like to predict. An elegant learning scheme for this prediction setting is the Structured Perceptron algorithm, which is sure to converge under some linear separability conditions. The framework integrates a very simple Structured layer on top of a latent costs network. Our key contribution is a novel loss function that incorporates structural information and simplifies learning. The effectiveness of this framework is illustrated with sequence prediction problems. We explore LSTM neural network architectures to model the latent costs layer, since our experiments concern NLP tasks. We perform basic experiments with Chunking in English. The SPN predictor outperforms its CRF equivalent. Our initial findings strongly indicate that SPN is a versatile framework with a powerful learning strategy. |
Year | DOI | Venue |
|---|---|---|
2018 | 10.1109/SSCI.2018.8628625 | 2018 IEEE Symposium Series on Computational Intelligence (SSCI) |
Keywords | Field | DocType |
latent cost learning,flexible modeling framework,prediction setting,linear separability conditions,latent costs network,structural information,sequence prediction problems,LSTM neural network architectures,latent costs layer,disjoint intervals,structured layer,NLP tasks,structured prediction networks,trees,learning scheme,structured perceptron algorithm | Sequence prediction,Linear separability,Disjoint sets,Task analysis,Computer science,Structured prediction,Artificial intelligence,Chunking (psychology),Artificial neural network,Perceptron,Machine learning | Conference |
ISBN | Citations | PageRank |
978-1-5386-9277-6 | 0 | 0.34 |
References | Authors | |
5 | 2 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Ruy Luiz Milidiú | 1 | 192 | 20.18 |
| Rafael Rocha | 2 | 0 | 0.34 |