Learning Markov Networks With Arithmetic CircuitsLearning Markov Networks With Arithmetic CircuitsLowd, Daniel and Rooshenas, Amirmohammad2013

Paper summaryopenreviewThe paper presents a learning algorithm for learning the structure of Markov Networks by exploring the connection between Arithmetic Circuits and MNs. This allows them to employ their previous work on learning ACs to learn a sub-set of MNs.
The main theme in the paper is to use “inference complexity as a learning bias.” Typically, in Markov network structure learning, one measures inference complexity using the number of edges added to the graph. In the present paper, the authors propose a more fine-grained measure: the number of edges added to an arithmetic circuit. As the authors show, this fine-grained measure yields an order of magnitude improvement in accuracy on several interesting benchmarks (I like the empirical evaluation).

The paper presents a learning algorithm for learning the structure of Markov Networks by exploring the connection between Arithmetic Circuits and MNs. This allows them to employ their previous work on learning ACs to learn a sub-set of MNs.
The main theme in the paper is to use “inference complexity as a learning bias.” Typically, in Markov network structure learning, one measures inference complexity using the number of edges added to the graph. In the present paper, the authors propose a more fine-grained measure: the number of edges added to an arithmetic circuit. As the authors show, this fine-grained measure yields an order of magnitude improvement in accuracy on several interesting benchmarks (I like the empirical evaluation).

Your comment:

You must log in before you can post this comment!

You must log in before you can submit this summary! Your draft will not be saved!

Preview:

0

Short Science allows researchers to publish paper summaries that are voted on and ranked! About