Conditional random field

A conditional random field (CRF) is a statistical modelling method often applied in pattern recognition. More specifically it is a type of discriminative undirected probabilistic graphical model. It is used to encode known relationships between observations and construct consistent interpretations. It is often used for labeling or parsing of sequential data, such as natural language text or biological sequences^[1] and in computer vision^[2] . Specifically, CRFs find applications in shallow parsing^[3] , named entity recognition^[4] and gene finding, among other tasks, being an alternative to the related hidden Markov models. In computer vision, CRFs are often used for object recognition and image segmentation.

Description

Lafferty, McCallum and Pereira (2001) define a CRF on observations $\boldsymbol{X}$ and random variables $\boldsymbol{Y}$ as follows:

Let G = (V,E) be a graph such that $\boldsymbol{Y} = (\boldsymbol{Y}_v)_{v\in V}$, so that $\boldsymbol{Y}$ is indexed by the vertices of G. Then $(\boldsymbol{X}, \boldsymbol{Y})$ is a conditional random field in case, when conditioned on $\boldsymbol{X}$, the random variables $\boldsymbol{Y}_v$ obey the Markov property with respect to the graph: $p(\boldsymbol{Y}_v |\boldsymbol{X}, \boldsymbol{Y}_w, w \neq v) = p(\boldsymbol{Y}_v |\boldsymbol{X}, \boldsymbol{Y}_w, w \sim v)$, where w∼v means that w and v are neighbors in G.

What this means is that a CRF is an undirected graphical model whose nodes can be divided into exactly two disjoint sets $\boldsymbol{X}$ and $\boldsymbol{Y}$, the observed and output variables, respectively; the conditional distribution $p(\boldsymbol{Y}|\boldsymbol{X})$ is then modeled.

Inference

For general graphs, the problem of exact inference in CRFs is intractable. The inference problem for a CRF is basically the same as for an MRF and the same arguments hold.^[5] However there exist special cases for which exact inference is feasible:

• If the graph is a chain or a tree, message passing algorithms yield exact solutions. The algorithms used in these cases are analogous to the forward-backward and Viterbi algorithm for the case of HMMs.
• If the CRF only contains pair-wise potentials and the energy is submodular, combinatorial min cut/max flow algorithms yield exact solutions.

If exact inference is impossible, several algorithms can be used to obtain approximate solutions. These include:

• Loopy belief propagation
• Alpha expansion
• Mean field inference
• Linear programming relaxations

Parameter Learning

Learning the parameters θ is usually done by maximum likelihood learning for p(Y_i | X_i;θ). If all nodes have exponential family distributions and all nodes are observed during training, this optimization is convex.^[5] It can be solved for example using gradient descent algorithms Quasi-Newton methods, such as the L-BFGS algorithm. On the other hand, if some variables are unobserved, the inference problem has to be solved for these variables. This is intractable to do exact in general graphs, so approximations have to be used.

Examples

In sequence modeling, the graph of interest is usually a chain graph. An input sequence of observed variables X represents a sequence of observations and Y represents a hidden (or unknown) state variable that needs to be inferred given the observations. The Y_i are structured to form a chain, with an edge between each Y_{i − 1} and Y_i. As well as having a simple interpretation of the Y_i as "labels" for each element in the input sequence, this layout admits efficient algorithms for:

• model training, learning the conditional distributions between the Y_i and feature functions from some corpus of training data.
• inference, determining the probability of a given label sequence Y given X.
• decoding, determining the most likely label sequence Y given X.

The conditional dependency of each Y_i on X is defined through a fixed set of feature functions of the form f(i,Y_{i − 1},Y_i,X), which can informally be thought of as measurements on the input sequence that partially determine the likelihood of each possible value for Y_i. The model assigns each feature a numerical weight and combines them to determine the probability of a certain value for Y_i.

Linear-chain CRFs have many of the same applications as conceptually simpler hidden Markov models (HMMs), but relax certain assumptions about the input and output sequence distributions. An HMM can loosely be understood as a CRF with very specific feature functions that use constant probabilities to model state transitions and emissions. Conversely, a CRF can loosely be understood as a generalization of an HMM that makes the constant transition probabilities into arbitrary functions that vary across the positions in the sequence of hidden states, depending on the input sequence.

Notably in contrast to HMMs, CRFs can contain any number of feature functions, the feature functions can inspect the entire input sequence X at any point during inference, and the range of the feature functions need not have a probabilistic interpretation.

Higher-order CRFs and semi-Markov CRFs

CRFs can be extended into higher order models by making each Y_i dependent on a fixed number o of previous variables Y_{i − o},...,Y_{i − 1}. Training and inference are only practical for small values of o (such as o ≤ 5),^{[citation needed]} since their computational cost increases exponentially with o. Large-margin models for structured prediction, such as the structured Support Vector Machine can be seen as an alternative training procedure to CRFs.

There exists another generalization of CRFs, the semi-Markov conditional random field (semi-CRF), which models variable-length segmentations of the label sequence Y.^[6] This provides much of the power of higher-order CRFs to model long-range dependencies of the Y_i, at a reasonable computational cost.

Software

This is a partial list of software that implement generic CRF tools.

• GCO CRFs with submodular energy functions (C++, Matlab)
• GRMM General CRFs (Java)
• CRFall General CRFs (Matlab)
• Sarawagi's CRF Linear-chain CRFs (Java)
• HCRF library Hidden-state CRFs (C++, Matlab)
• Wapiti Fast linear-chain CRFs (C)
• CRFSuite Fast restricted linear-chain CRFs (C)
• CRF++ Linear-chain CRFs (C++)
• Monte Python Linear-chain CRFs (Python)

This is a partial list of software that implement CRF related tools.

• Conrad CRF based gene predictor (Java)
• Stanford NER Named Entity Recognizer (Java)
• BANNER Named Entity Recognizer (Java)

See also

• Graphical model
• Markov random field
• Maximum entropy Markov model (MEMM)

References

1. ^ Lafferty, J., McCallum, A., Pereira, F. (2001). "Conditional random fields: Probabilistic models for segmenting and labeling sequence data". Proc. 18th International Conf. on Machine Learning. Morgan Kaufmann. pp. 282–289. http://www.cis.upenn.edu/~pereira/papers/crf.pdf. 
2. ^ He, X. and Zemel, R.S. and Carreira-Perpinñán, M.A. (2004). "Multiscale conditional random fields for image labeling". IEEE Computer Society. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.3.7826&rep=rep1&type=pdf. 
3. ^ Sha, F., Pereira, F. (2003). "shallow parsing with conditional random fields". http://portal.acm.org/ft_gateway.cfm?id=1073473&type=pdf&CFID=4684435&CFTOKEN=39459323. 
4. ^ Settles, B. (2004). "Biomedical named entity recognition using conditional random fields and rich feature sets". Proceedings of the International Joint Workshop on Natural Language Processing in Biomedicine and its Applications. pp. 104--107. http://acl.ldc.upenn.edu/coling2004/W1/pdf/21.pdf. 
5. ^ ^{a b} Sutton, Charles; McCallum, Andrew (2010). "An Introduction to Conditional Random Fields". arXiv:1011.4088v1 [stat.ML]. 
6. ^ Sarawagi, Sunita; William W. Cohen (2005). "Semi-Markov conditional random fields for information extraction". In Lawrence K. Saul, Yair Weiss, Léon Bottou (eds.). Advances in Neural Information Processing Systems 17. Cambridge, MA: MIT Press. pp. 1185–1192. 

Further reading

• McCallum, A.: Efficiently inducing features of conditional random fields. In: Proc. 19th Conference on Uncertainty in Artificial Intelligence. (2003)

• Wallach, H.M.: Conditional random fields: An introduction. Technical report MS-CIS-04-21, University of Pennsylvania (2004)
• Sutton, C., McCallum, A.: An Introduction to Conditional Random Fields for Relational Learning. In "Introduction to Statistical Relational Learning". Edited by Lise Getoor and Ben Taskar. MIT Press. (2006) Online PDF
• Klinger, R., Tomanek, K.: Classical Probabilistic Models and Conditional Random Fields. Algorithm Engineering Report TR07-2-013, Department of Computer Science, Dortmund University of Technology, December 2007. ISSN 1864-4503. Online PDF