Yoshifumi Sakai

Fresh Information Retrieval Using Cooperative Meta Search Engines

In web page retrievals, search engines are usually used. However, conventional search engines have a problem in that their update intervals are very longb ecause they are based on centralized architecture, which gathers documents using robots. So we proposed the Cooperative Search Engine (CSE) in order to reduce the update interval. CSE is a distributed search engine, which integrates small local search engines into a large global search engine by using local meta search engines. A local meta search engine hides a local search engine in each web site. Although CSE can reduce the update interval, the retrieval performance is not enough. So, we proposed several speed up techniques. In this paper, we describe the structure and behavior of CSE and its efficiency.

Persistent cache in Cooperative Search Engine

Cooperative Search Engine (CSE) is a distributed search engine, which can update indexes in a very short time for the purpose of fresh information retrieval. In CSE, the retrieval performance is dependent on cache contents because communication delay occurs at retrieval time. On the other hand, however, cache is invalidated as soon as indexes are updated. Therefore, we need persistent cache that can hold valid data before and after updating. In this paper we describe the principle and evaluations of persistent cache.

Learning Monotone Log-Term DNF Formulas under the Uniform Distribution

Abstract. Based on the uniform distribution PAC learning model, the learnability for the class of monotone disjunctive normal form formulas with at most O (log n ) terms, denoted O (log n )-term MDNF, is investigated. Using the technique of restriction, an algorithm that learns O (log n )-term MDNF by examples in polynomial time is given.

Distributed information retrieval by using cooperative meta search engines

Information retrieval is frequently used for the Internet. However, we face the problem that the update intervals of conventional search engines are too long. It is difficult for conventional search engines to make their intervals short because they are based on a centralized architecture. So, we proposed the Cooperative Search Engine (CSE), which is based on a distributed architecture. CSE can realize the shortening of the update interval. In CSE, a large search engine is constructed from multiple local meta search engines that cooperate with each other. CSE can accomplish correct ranking in a distributed fashion because each local meta search engine hides the difference between its local search engine and the others. The article discusses the design and experimental evaluations of CSE.

In silico analysis of 3'-end-processing signals in Aspergillus oryzae using expressed sequence tags and genomic sequencing data.

To investigate 3′-end-processing signals in Aspergillus oryzae, we created a nucleotide sequence data set of the 3′-untranslated region (3′ UTR) plus 100 nucleotides (nt) sequence downstream of the poly(A) site using A. oryzae expressed sequence tags and genomic sequencing data. This data set comprised 1065 sequences derived from 1042 unique genes. The average 3′ UTR length in A. oryzae was 241 nt, which is greater than that in yeast but similar to that in plants. The 3′ UTR and 100 nt sequence downstream of the poly(A) site is notably U-rich, while the region located 15–30 nt upstream of the poly(A) site is markedly A-rich. The most frequently found hexanucleotide in this A-rich region is AAUGAA, although this sequence accounts for only 6% of all transcripts. These data suggested that A. oryzae has no highly conserved sequence element equivalent to AAUAAA, a mammalian polyadenylation signal. We identified that putative 3′-end-processing signals in A. oryzae, while less well conserved than those in mammals, comprised four sequence elements: the furthest upstream U-rich element, A-rich sequence, cleavage site, and downstream U-rich element flanking the cleavage site. Although these putative 3′-end-processing signals are similar to those in yeast and plants, some notable differences exist between them.

A linear space algorithm for computing a longest common increasing subsequence

Let X and Y be sequences of integers. A common increasing subsequence of X and Y is an increasing subsequence common to X and Y. In this note, we propose an O(|X| ċ |Y|)-time and O(|x| + |Y|)-space algorithm for finding one of the longest common increasing subsequences of X and Y, which improves the space complexity of Yang et al. [I.H. Yang, C.P. Huang, K.M. Chao, A fast algorithm for computing a longest common increasing subsequence, Inform. Process. Lett. 93 (2005) 249-253] O(|X| ċ |Y|)-time and O(|X| ċ |Y|)-space algorithm, where |X| and |Y| denote the lengths of X and Y, respectively.

Temporal information retrieval in cooperative search engine

In business, the retrieval of up-to-data, or fresh, information is very important. It is difficult for conventional search engines based on a centralized architecture to retrieve fresh information, because they take a long time to collect documents via Web robots. In contrast to a centralized architecture, a search engine based on a distributed architecture does not need to collect documents, because each site makes an index independently. As a result, distributed search engines can be used to retrieve fresh information. However, fast indexing alone is not enough to retrieve fresh information, as support fro temporal information based retrieval is also required. In this paper, we describe temporal information retrieval in distributed search engines. In particular, we propose a content-based comparison method to avoid spamming.

Learning monotone log-term DNF formulas

Based on the uniform distribution PAC learning model, the learnability for monotone disjunctive normal form formulas with at most <italic>O</italic>(log<italic>n</italic>) terms (<italic>O</italic>(log<italic>n</italic>)-term MDNF) is investigated. Using the technique of restriction, an algorithm that learns <italic>O</italic>(log<italic>n</italic>)-term MDNF in polynomial time is given.

A distributed search engine for fresh information retrieval

In information retrieval of web pages, we usually use search engines. A conventional search engine has the problem that the update interval is very long because it is based on centralized architecture, which collects documents by using robots. So, we have developed a distributed search engine, cooperative search engine (CSE) in order to reduce the update interval. In CSE, a global search engine is constructed with multiple meta search engines, which hide each local search engine in each Web site. CSE can make the update interval very short by indexing on each web site. In this paper we describe the framework of CSE and its efficiency.

Proper learning algorithm for functions of k terms under smooth distributions

Algorithms for learning feasibly Boolean functions from examples are explored. A class of functions we deal with is written as F1 o~z~ = {v(fl(u),... )fk~’)))) 19 E