Generalized Conditional Random Fields
نویسنده
چکیده
Conditional random fields (CRFs) have shown significant improvements over existing methods for structured data labeling. However independence assumptions made by CRFs decrease the usability of the models produced. Currently, CRF models accomodate dependence between only adjacent labels. Generalized CRFs proposed in this study relaxes assumptions of CRFs without reducing tractability of inference and training. Since constraints can be generalized arbitrarily in the proposed framework, linear-chain CRFs become the simplest case of generalized CRFs. Although relaxed independence assumptions require additional overhead, arbitrarily large dependencies introduce only polynomial computational complexity. Experimental results confirmed that the proposed framework increases both accuracy and applicability of CRFs.
منابع مشابه
Generalized Expectation Criteria for Semi-Supervised Learning of Conditional Random Fields
This paper presents a semi-supervised training method for linear-chain conditional random fields that makes use of labeled features rather than labeled instances. This is accomplished by using generalized expectation criteria to express a preference for parameter settings in which the model’s distribution on unlabeled data matches a target distribution. We induce target conditional probability ...
متن کاملTrajectory Learning Based on Conditional Random Fields for Robot Programming by Demonstration
This work presents an approach for implementation of conditional random fields (CRF) in transferring motor skills to robots. As a discriminative probabilistic model, CRF models directly the conditional probability distribution over label sequences for given observation sequences. Hereby, CRF was employed for segmentation and labeling of a set of demonstrated trajectories observed by a tracking ...
متن کاملConditional Random Fields for Airborne Lidar Point Cloud Classification in Urban Area
Over the past decades, urban growth has been known as a worldwide phenomenon that includes widening process and expanding pattern. While the cities are changing rapidly, their quantitative analysis as well as decision making in urban planning can benefit from two-dimensional (2D) and three-dimensional (3D) digital models. The recent developments in imaging and non-imaging sensor technologies, s...
متن کاملMinimum Classification Error Training of Hidden Conditional Random Fields for Speech and Speaker Recognition
Hidden conditional random fields (HCRFs) are derived from the theory of conditional random fields with hidden-state probabilistic framework. It directly models the conditional probability of a label sequence given observations. Compared to hidden Markov models, HCRFs provide a number of benefits in the acoustic modeling of speech signals. Prior works for training on HCRFs were accomplished with...
متن کاملGeneralized Expectation Criteria for Semi-Supervised Learning with Weakly Labeled Data
In this paper, we present an overview of generalized expectation criteria (GE), a simple, robust, scalable method for semi-supervised training using weakly-labeled data. GE fits model parameters by favoring models that match certain expectation constraints, such as marginal label distributions, on the unlabeled data. This paper shows how to apply generalized expectation criteria to two classes ...
متن کامل