Masaya Murata

Learn More
This paper reports our methods and experimental results for two TRECVID 2012 tasks of instance search and multimedia event detection. For the instance search, we applied the BM25(Best Match 25) method, the state-of-the-art ranking function in the field of text retrieval, to the video retrieval task. Our BM25 features the following three factors: (i) the(More)
Users of text retrieval systems input only a few keywords or sometimes just one keyword to the systems even if they had complex information needs. Due to the lack of query keywords, it becomes hard to return relevant search results that satisfy the demands of each user. Because digital documents, in contrast to queries, are generally composed of many kinds(More)
Information needs expressed by using the same query for a search engine might be totally different, whether on week days or weekends, or during the day or at night. For queries having no temporal changes in search intentions, the same search results ranking may be returned regardless of the time, but for those with temporal changes the ranking must be(More)
— A Kalman filter (KF) is state-of-the-art for estimating states of linear-Gaussian state-space models. The KF selects an expectation of a posterior probability density function of state and the expectation is an analytic solution for minimizing the square estimation error. The estimate of KF is therefore optimal, however, simultaneously inherits the(More)
This paper reports our method and experimental result on the TRECVID 2014[1] instance search task. Since 2012, we have been applying BM25 (Best Match 25), i.e., the state-of-the-art probabilistic information retrieval method in the field of text retrieval, to the instance search task. The standard BM25 uses the well-known Inverse Document Frequency (IDF) as(More)
This paper proposes to use a radial basis function (RBF) network to increase the separation performance of blind signal separation (BSS). Independent component analysis (ICA) is often used for BSS, but in general, ICA employs a sigmoid function to describe the probability distribution of signals in the process of learning. We attempt to describe the(More)