Name
Abstract | ||
|---|---|---|
Many real-world classification tasks involve the prediction of multiple, inter-dependent class labels. A prototypical case of this sort deals with prediction of a sequence of labels for a sequence of observations. Such problems arise naturally in the context of annotating and segmenting observation sequences. This paper generalizes Gaussian Process classification to predict multiple labels by taking dependencies between neighboring labels into account. Our approach is motivated by the desire to retain rigorous probabilistic semantics, while overcoming limitations of parametric methods like Conditional Random Fields, which exhibit conceptual and computational difficulties in high-dimensional input spaces. Experiments on named entity recognition and pitch accent prediction tasks demonstrate the competitiveness of our approach. |
Year | DOI | Venue |
|---|---|---|
2004 | 10.1145/1015330.1015433 | ICML |
Keywords | Field | DocType |
real-world classification task,multiple label,gaussian process classification,entity recognition,pitch accent prediction task,high-dimensional input space,inter-dependent class label,observation sequence,annotating sequence,computational difficulty,conditional random fields,conditional random field,gaussian process | Conditional random field,Parametric methods,Market segmentation,Pattern recognition,Computer science,sort,Pitch accent,Gaussian process,Artificial intelligence,Named-entity recognition,Machine learning,Probabilistic semantics | Conference |
ISBN | Citations | PageRank |
1-58113-838-5 | 18 | 0.98 |
References | Authors | |
18 | 3 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| yasemin altun | 1 | 2463 | 150.46 |
| Thomas Hofmann | 2 | 10064 | 1001.83 |
| Alexander J. Smola | 3 | 19627 | 1967.09 |