Branimir Todorovic

Learn More
The problem of blind source separation consists in retrieving unknown source signals from their mixtures. In the estimation of sources, most of the proposed algorithms use stochastic or natural gradient to optimize some contrast function obtained from the independence property of sources. In this paper, we deal with non-stationary source signals, and(More)
We present a primal sub-gradient method for structured SVM optimization defined with the averaged sum of hinge losses inside each example. Compared with the mini-batch version of the Pegasos algorithm for the structured case, which deals with a single structure from each of multiple examples, our algorithm considers multiple structures from a single example(More)
—The paper proposes a new message passing algorithm for cycle-free factor graphs. The proposed " entropy message passing " (EMP) algorithm may be viewed as sum-product message passing over the entropy semiring, which has previously appeared in automata theory. The primary use of EMP is to compute the entropy of a model. However, EMP can also be used to(More)
Corefence resolution is the process of determining whether two expressions in natural language refer to the same entity in the world. We adopt machine learning approach using decision tree to a coreference resolution of general noun phrases in unrestricted text based on well defined features. We follow the work of Soon et al. [25] and demonstrate the(More)
Structured learning algorithms usually require inference during the training procedure. Due to their exponential size of output space, the parameter update is performed only on a relatively small collection built from the " best " structures. The k-best MIRA is an example of an online algorithm which seeks optimal parameters by making updates on k(More)
The paper proposes a new recursive algorithm for the exact computation of the linear chain conditional random fields gradient. The algorithm is an instance of the Entropy Message Passing (EMP), introduced in our previous work, and has the purpose to enhance memory efficiency when applied to long observation sequences. Unlike the traditional algorithm based(More)