• Journal of Information Processing Systems
• /
• 제19권3호
• /
• pp.355-369
• /
• 2023

Since the widespread adoption of deep-learning and related distributed representation, there have been substantial advancements in part-of-speech (POS) tagging for many languages. When training word representations, morphology and shape are typically ignored, as these representations rely primarily on collecting syntactic and semantic aspects of words. However, for tasks like POS tagging, notably in morphologically rich and resource-limited language environments, the intra-word information is essential. In this study, we introduce a deep neural network (DNN) for POS tagging that learns character-level word representations and combines them with general word representations. Using the proposed approach and omitting hand-crafted features, we achieve 90.47%, 80.16%, and 79.32% accuracy on our own dataset for three morphologically rich languages: Uyghur, Uzbek, and Kyrgyz. The experimental results reveal that the presented character-based strategy greatly improves POS tagging performance for several morphologically rich languages (MRL) where character information is significant. Furthermore, when compared to the previously reported state-of-the-art POS tagging results for Turkish on the METU Turkish Treebank dataset, the proposed approach improved on the prior work slightly. As a result, the experimental results indicate that character-based representations outperform word-level representations for MRL performance. Our technique is also robust towards the-out-of-vocabulary issues and performs better on manually edited text.

• Asian Pacific Journal of Cancer Prevention
• /
• 제14권12호
• /
• pp.7459-7466
• /
• 2013

Background: Some 60 years after introduction of the Papanicolaou smear worldwide, cervical cancer remains a burden in developing countries where >85% of world new cases and deaths occur, suggesting a failure to establish comprehensive cervical-cancer control programs. Effective interventions are available to control cervical cancer but are not all affordable in low-income settings. Disease awareness saves lives by risk-reduction as witnessed in reducing mortality of HIV/AIDS and smoking-related cancers. Subjects and Methods: We initiated a community-based awareness program on cervical cancer in two low-income Muslim Uyghur townships in Kashi (Kashgar) Prefecture, Xinjiang, China in 2008. The education involved more than 5,000 women from two rural townships and awareness was then evaluated in 2010 and 2011, respectively, using a questionnaire with 10 basic knowledge questions on cervical cancer. Demographic information was also collected and included in an EpiData database. A 10-point scoring system was used to score the awareness. Results: The effectiveness and feasibility of the program were evaluated among 4,475 women aged 19-70 years, of whom >92% lived on/below US$1.00/day. Women without prior education showed a poor average awareness rate of 6.4% (164/2,559). A onetime education intervention, however, sharply raised the awareness rate by 4-fold to 25.5% (493/1,916). Importantly, low income and illiteracy were two reliable factors affecting awareness before or after education intervention. Conclusions: Education intervention can significantly raise the awareness of cervical cancer in low-income women. Economic development and compulsory education are two important solutions in raising general disease awareness. We propose that implementing community-based awareness programs against cervical cancer is realistic, locally affordable and sustainable in low-income countries, which may save many lives over time and, importantly, will facilitate the integration of comprehensive programs when feasible. In this context, adopting this strategy may provide one good example of how to achieve "good health at low cost".