Many Japanese words are made of kanji characters, which themselves represent meanings. However traditional word-based distributional semantic models (DSMs) do not benefit from the useful semantic information of kanji characters. In this paper, we propose a method for exploiting the semantic information of kanji characters for constructing Japanese word vectors in DSMs. In the proposed method, t...

This paper explores an issue of redundant errors reported while automatically scoring English learners’ sentences. We use a human-computer collaboration approach to eliminate redundant errors. The first step is to automatically select candidate redundant errors using PMI and RFC. Since those errors are detected with different IDs although they represent the same error, the candidacy cannot be c...

This paper investigates the availability characteristics and the reliability of a three-dissimilar-unit repairable system with two diierent repair facilities. Under some practical assumptions, we obtain the explicit expressions of the state probabilities of the system, and then the explicit expressions of the following performance measures of the system: (1) the pointwise and steady-state avail...

In this paper, we study the problem of pointwise estimation of a multivariate function. We develop a general pointwise estimation procedure that is based on selection of estimators from a large parameterized collection. An upper bound on the pointwise risk is established and it is shown that the proposed selection procedure specialized for different collections of estimators leads to minimax an...

We consider pointwise, box, and Hausdorff dimensions of invariant measures for circle diffeomorphisms. We discuss the cases of rational, Diophantine, and Liouville rotation numbers. Our main result is that for any Liouville number τ there exists a C∞ circle diffeomorphism with rotation number τ such that the pointwise and box dimensions of its unique invariant measure do not exist. Moreover, th...

Recent exploratory efforts in discourse-level language modeling have relied heavily on calculating Pointwise Mutual Information (PMI), which involves significant computation when done over large collections. Prior work has required aggressive pruning or independence assumptions to compute scores on large collections. We show the method of Conditional Random Sampling, thus far an underutilized t...