Search-based structured prediction

  title={Search-based structured prediction},
  author={Hal Daum{\'e} and J. Langford and D. Marcu},
  journal={Machine Learning},
  • Hal Daumé, J. Langford, D. Marcu
  • Published 2009
  • Mathematics, Computer Science
  • Machine Learning
  • We present Searn, an algorithm for integrating search and learning to solve complex structured prediction problems such as those that occur in natural language, speech, computational biology, and vision. Searn is a meta-algorithm that transforms these complex problems into simple classification problems to which any binary classifier may be applied. Unlike current algorithms for structured learning that require decomposition of both the loss function and the feature functions over the predicted… CONTINUE READING
    518 Citations

    Figures, Tables, and Topics from this paper

    Unsupervised search-based structured prediction
    • 34
    • PDF
    Inference algorithms for the regression approach to sequence prediction
    • PDF
    Output Space Search for Structured Prediction
    • 20
    • PDF
    Span-based Neural Structured Prediction
    Learning to Speed Up Structured Output Prediction
    • 3
    • PDF
    Learning Constraints for Structured Prediction Using Rectifier Networks
    • PDF
    Randomized Greedy Search for Structured Prediction: Amortized Inference and Learning
    • 1
    • Highly Influenced
    • PDF
    ℋC-search for structured prediction in computer vision
    • 29
    • PDF


    Practical structured learning techniques for natural language processing
    • 112
    • PDF
    Learning structured prediction models: a large margin approach
    • 528
    • PDF
    Large Margin Methods for Structured and Interdependent Output Variables
    • 2,186
    • PDF
    Gaussian process classification for segmenting and annotating sequences
    • 78
    • PDF
    A Framework for Learning Predictive Structures from Multiple Tasks and Unlabeled Data
    • 1,282
    • PDF
    Discriminative Sentence Compression with Soft Syntactic Evidence
    • 171
    • PDF