三个IT厂商酿成的惨痛教训

ｋｅｙｓｔｏｎｅ公司长期停滞的ＥＲＰ项目终于宣告成功，　系统将 １３０ 个厂房，全美国２０００ 多员工连接在一起。回首ｋｅｙｓｔｏｎｅ 信息化历程，其经验教训值得我们深思。     

One hundred and fifty years of sprint and distance running – Past trends and future prospects

Sprint and distance running have experienced remarkable performance improvements over the past century. Attempts to forecast running performances share an almost similarly long history but have relied so far on relatively short data series. Here, we compile a comprehensive set of season-best performances for eight Olympically contested running events. With this data set, we conduct (1) an exponential time series analysis and (2) a power-law experience curve analysis to quantify the rate of past performance improvements and to forecast future performances until the year 2100. We find that the sprint and distance running performances of women and men improve exponentially with time and converge at yearly rates of 4% ± 3% and 2% ± 2%, respectively, towards their asymptotic limits. Running performances can also be modelled with the experience curve approach, yielding learning rates of 3% ± 1% and 6% ± 2% for the women's and men's events, respectively. Long-term trends suggest that: (1) women will continue to run 10–20% slower than men, (2) 9.50 s over 100 m dash may only be broken at the end of this century and (3) several middle- and long-distance records may be broken within the next two to three decades. The prospects of witnessing a sub-2 hour marathon before 2100 remain inconclusive. Our results should be interpreted cautiously as forecasting human behaviour is intrinsically uncertain. The future season-best sprint and distance running performances will continue to scatter around the trends identified here and may yield unexpected improvements of standing world records.     

Authorship attribution based on a probabilistic topic model

This paper describes, evaluates and compares the use of Latent Dirichlet allocation (LDA) as an approach to authorship attribution. Based on this generative probabilistic topic model, we can model each document as a mixture of topic distributions with each topic specifying a distribution over words. Based on author profiles (aggregation of all texts written by the same writer) we suggest computing the distance with a disputed text to determine its possible writer. This distance is based on the difference between the two topic distributions. To evaluate different attribution schemes, we carried out an experiment based on 5408 newspaper articles (Glasgow Herald) written by 20 distinct authors. To complement this experiment, we used 4326 articles extracted from the Italian newspaper La Stampa and written by 20 journalists. This research demonstrates that the LDA-based classification scheme tends to outperform the Delta rule, and the χ² distance, two classical approaches in authorship attribution based on a restricted number of terms. Compared to the Kullback–Leibler divergence, the LDA-based scheme can provide better effectiveness when considering a larger number of terms.     

Searching strategies for the Hungarian language

This paper reports on the underlying IR problems encountered when dealing with the complex morphology and compound constructions found in the Hungarian language. It describes evaluations carried out on two general stemming strategies for this language, and also demonstrates that a light stemming approach could be quite effective. Based on searches done on the CLEF test collection, we find that a more aggressive suffix-stripping approach may produce better MAP. When compared to an IR scheme without stemming or one based on only a light stemmer, we find the differences to be statistically significant. When compared with probabilistic, vector-space and language models, we find that the Okapi model results in the best retrieval effectiveness. The resulting MAP is found to be about 35% better than the classical tf idf approach, particularly for very short requests. Finally, we demonstrate that applying an automatic decompounding procedure for both queries and documents significantly improves IR performance (+10%), compared to word-based indexing strategies.     

A true original

Abstract

This study analyzes Japanese and Korean ethno-national (minzoku-kokumin) education in postwar Japan. During a period of political unrest in Japan (1945–1955), some of the Korean residents and Japanese worked together to overcome the culture of Imperial Japan and its assimilative education. They also regarded themselves as people colonized by the United States, and pursued a political-cultural movement for their liberation and independence from American imperialism. The Koreans in Japan rejected compulsory education in the Japanese language. As a result, since 1956, Korean schools (Chōsengakkō), funded and supported by North Korea, were founded all across Japan. Their ethno-national education was in fact incorporated into North Korean politics, and has been considered in many studies as having overcome Japanese assimilation and ethnic inequalities. Such a view was a result of many academic Zainichi Korean studies that come from an “insider's perspective” to criticize Japanese colonialism and discrimination. In order to go beyond this insider's view, I focus on the political alliance between Zainichi Koreans and the Japanese people in their pursuit of ethno-national education. Since 2010, the Japanese state funding for Korean schools has become a major controversy in Japan. By tracing the historical background, this article intends to explain why this political issue has arisen. The ultimate purpose of this article is to suggest an ethical perspective to resolve the current political conflict regarding Korean schools in Japan.