Complex Predictor Attribute

A Complex Predictor Attribute is a predictor attribute that performs a complex transformation (that is derived from on lower-level predictor features).

• Context:
  • It can range from being a Dense Complex Predictor Attribute to being a Sparse Complex Predictor Attribute.
  • It can be associated to a Feature Detector (in a NNet).
  • …
• Example(s):
  • a Global Predictor Feature.
  • a Graph Path-based Predictor Feature.
  • an Image Line Feature.
  • a Predicted Word Part-of-Speech Role Feature.
  • a Temporal Summary Feature.
  • …
• Counter-Example(s):
  • Low-level Predictor Feature.
• See: Higher-Level Predictor Feature Learning, Feature Extraction.

---

---

References

2009a

• (Ye et al., 2009) ⇒ Nan Ye, Wee Sun Lee, Hai Leong Chieu, and Dan Wu. (2009). “Conditional Random Fields with High-Order Features for Sequence Labeling.” In: Advances in Neural Information Processing Systems 22 (NIPS 2009)
  • QUOTE: we show that it is possible to design efficient inference algorithms for a conditional random field using features that depend on long consecutive label sequences (high-order features), as long as the number of distinct label sequences used in the features is small.

2009b

• (Qian et al., 2009) ⇒ Xian Qian, Xiaoqian Jiang, Qi Zhang, Xuanjing Huang, and Lide Wu. (2009). “Sparse Higher Order Conditional Random Fields for Improved Sequence Labeling.” In: Proceedings of ICML Conference (ICML 2009).
  • QUOTE: In real sequence labeling tasks, statistics of many higher order features are not sufficient due to the training data sparseness, very few of them are useful. We describe Sparse Higher Order Conditional Random Fields (SHO-CRFs), which are able to handle local features and sparse higher order features together using a novel tractable exact inference algorithm.

2007a

• (Hinton, 2007) ⇒ Geoffrey E. Hinton. (2007). “Learning Multiple Layers of Representation.” In: Trends in cognitive sciences, 11(10).
  • QUOTE: … level representations, and bottom-up connections can be used to infer the high-level representations that … modeled by using the hidden layer of an RBM to model the higher-order correlations between … recognition weights are used to infer binary states for the 500 feature units …

2007b

• (Choras, 2007) ⇒ Ryszard S. Choras. (2007). “Image Feature Extraction Techniques and their Applications for CBIR and Biometrics Systems.” In: International journal of biology and biomedical engineering, 1(1).
  • QUOTE: … The feature is defined as a function of one or more measurements, each of which specifies some quantifiable property of an object, and is computed such that … Low-level features can be extracted directed from the original images, whereas high-level feature extraction must …

2006

• (McDonald & Pereira, 2006) ⇒ Ryan T. McDonald, and Fernando Pereira. (2006). “Online Learning of Approximate Dependency Parsing Algorithms.” In: Proceedings of EACL (EACL 2006).
  • QUOTE: In this paper we extend the MST parsing framework to incorporate higher-order feature representations of bounded-size connected subgraphs.

---