著者
芦川 将之 川村 隆浩 大須賀 昭彦
出版者
一般社団法人 人工知能学会
雑誌
人工知能学会論文誌 (ISSN:13460714)
巻号頁・発行日
vol.32, no.3, pp.B-G81_1-13, 2017-05-01 (Released:2017-05-01)
参考文献数
31

Current crowdsourcing platforms such as Amazon Mechanical Turk provide an attractive solution Crowdsourcing platforms provide an attractive solution for processing numerous tasks at a low cost. However, insufficient quality control remains a major concern. Therefore, we developed a private crowdsourcing system that allows us to devise quality control methods. In the present study, we propose a grade-based training method for workers in order to avoid simple exclusion of low-quality workers and shrinkage of the crowdsourcing market in the near future. Our training method utilizes probabilistic networks to estimate correlations between tasks based on workers’ records for 18.5 million tasks and then allocates pre-learning tasks to the workers to raise the accuracy of target tasks according to the task correlations. In an experiment, the method automatically allocated 31 pre-learning task categories for 9 target task categories, and after the training of the pre-learning tasks, we confirmed that the accuracy of the target tasks was raised by 7.8 points on average. This result was comparatively higher than those of pre-learning tasks allocated using other methods, such as decision trees. We thus confirmed that the task correlations can be estimated using a large amount of worker records, and that these are useful for the grade-based training of low-quality workers.
著者
芦川 将之 川村 隆浩 大須賀 昭彦
出版者
一般社団法人 人工知能学会
雑誌
人工知能学会論文誌 (ISSN:13460714)
巻号頁・発行日
vol.29, no.6, pp.503-515, 2014-11-01 (Released:2014-09-25)
参考文献数
25

Open Crowdsourcing platforms like Amazon Mechanical Turk provide an attractive solution for process of high volume tasks with low costs. However problems of quality control is still of major interest. In this paper, we design a private crowdsourcing system, where we can devise methods for the quality control. For the quality control, we introduce four worker selection methods, each of which we call preprocessing filtering, real-time filtering, post processing filtering, and guess processing filtering. These methods include a novel approach, which utilizes a collaborative filtering technique in addition to a basic approach of initial training or gold standard data. For an use case, we have built a very large dictionary, which is necessary for Large Vocabulary Continuous Speech Recognition and Text-to-Speech. We show how the system yields high quality results for some difficult tasks of word extraction, part-of-speech tagging, and pronunciation prediction to build a large dictionary.
著者
芦川 将之 川村 隆浩 大須賀 昭彦
出版者
The Japanese Society for Artificial Intelligence
雑誌
人工知能学会論文誌 (ISSN:13460714)
巻号頁・発行日
vol.29, no.6, pp.503-515, 2014

Open Crowdsourcing platforms like Amazon Mechanical Turk provide an attractive solution for process of high volume tasks with low costs. However problems of quality control is still of major interest. In this paper, we design a private crowdsourcing system, where we can devise methods for the quality control. For the quality control, we introduce four worker selection methods, each of which we call preprocessing filtering, real-time filtering, post processing filtering, and guess processing filtering. These methods include a novel approach, which utilizes a collaborative filtering technique in addition to a basic approach of initial training or gold standard data. For an use case, we have built a very large dictionary, which is necessary for Large Vocabulary Continuous Speech Recognition and Text-to-Speech. We show how the system yields high quality results for some difficult tasks of word extraction, part-of-speech tagging, and pronunciation prediction to build a large dictionary.
著者
中田 康太 芦川 将之
出版者
人工知能学会
雑誌
人工知能学会全国大会論文集 (ISSN:13479881)
巻号頁・発行日
vol.27, 2013

音声データの収集は時間的・経済的コストが大きいことが知られている。本研究では、クラウドソーシングを利用した読み上げ音声収集により、安価かつ高速に音声データを収集する。また読み上げテキストと音響の情報を用いたフィルタを設計することで、正確な読み上げ音声データを自動抽出できることを示す。収集した音声の音声認識タスクへの応用や、ゲーミフィケーションの可能性についても報告する。