Search Results 1-20 of 619

  • Speech Recognition Error Identifier in Specific Domain Independent of Automatic Speech Recognition Engine  [in Japanese]

    田中 拓 , 野田 直希 , 加藤 恒夫 , 山本 誠一

    … However, generic ASR engines sometimes output erroneous results which do not fit the domain or context. … Although the verification of ASR results has been defined as a general sequence labeling task using internal variables of an ASR engine conventionally, we propose ASR error identifiers which estimate if a result contains errors or not independent of the ASR engine, trained with limited amount of log data of the domain. …

    情報処理学会論文誌 61(3), 608-617, 2020-03-15

    IPSJ 

  • Junior high school output activities based on the "Large Grammar" method ―― Using English as a communication tool.  [in Japanese]

    石田 順 , 金森 玲子 , 竹川 由紀子 , 足立 和美

    … At Tottori University junior high school we practice output activities on a daily basis in our English classes with the aim of enhancing students' ability to communicate in English. … These activities, using the Large Grammar method proposed by one of the authors, Kazumi Adachi, rely on improvised output. … We are aiming for activities that encourage students to think "voluntarily", "interactively" and "deeply" about the task. …

    鳥取大学附属中学校研究紀要 (51), 73-106, 2020-03-01

    IR 

  • Oral Output Training in EFL Context : The Effects of Task Repetition on Speech Complexity, Accuracy, and Fluency

    Iwata Akira

    Research bulletin of English teaching : JACET北海道支部紀要 (16), 1-27, 2020

  • Long-Term Prediction of Small Time-Series Data Using Generalized Distillation  [in Japanese]

    Hayashi Shogo , Tanimoto Akira , Kashima Hisashi

    … In particular, for long-term time-series prediction, thedata size tends to be small because of the unavailability of the data between input and output times in training. … The key idea of the proposed method is to utilize the middle-time data betweenthe input and output times as "privileged information," which is available only in the training phase and not in thetest phase. …

    Transactions of the Japanese Society for Artificial Intelligence 35(5), B-K33_1-9, 2020

    J-STAGE 

  • Policy Gradient Reinforcement Learning for Policy Represented by Fuzzy Rules: Application to Simulations on Speed Control of an Automobile

    ISHIHARA Seiji , IGARASHI Harukazu

    … A study has applied this method to a task of speed control of an automobile and has obtained correct policies, some of which control speed of the automobile appropriately but many others generate inappropriate vibration of speed. … In general, the policy is not desirable that causes sudden time change or vibration in the output value, and there would be many cases where the policy giving smooth time change in the output value is desirable. …

    Journal of Japan Society for Fuzzy Theory and Intelligent Informatics 32(4), 801-810, 2020

    J-STAGE 

  • A Study on Text-to-SQL with Data Values  [in Japanese]

    KAKU Soichiro , NISHIDA Kyosuke , TOMITA Junji

    … In previous techniques, a natural language question and a DB schema are used as inputs to output an SQL query. … In this study, we define a sub- task that determines whether or not a DB column is used in an SQL query to get the answer to a question using data values. … The column determination model learned in this task can be combined with a text-to-SQL model.</p> …

    Proceedings of the Annual Conference of JSAI JSAI2020(0), 4Rin109-4Rin109, 2020

    J-STAGE 

  • Dynamic Microtask Re-posting for Efficient Answer Collection in Crowdsourcing  [in Japanese]

    MORINAGA Masaya , SAITO Susumu , NAKANO Teppei , KOBAYASHI Tetsunori , OGAWA Tetsuji

    … <p>To improve the quality of the output in crowdsourcing at low cost, it is effective to change the number of collecting answers for each question and make a majority decision. … By using a binary classification task and simulating the dynamic collection of answers for development set with correct labels, determine the most efficient way to perform labeling. …

    Proceedings of the Annual Conference of JSAI JSAI2020(0), 4Rin174-4Rin174, 2020

    J-STAGE 

  • Open-Domain Question-Answering using NLG based query expansion and ranking method  [in Japanese]

    KOYANAGI Ryuto , YAMAMOTO Kotaro , SUZUKI Kentaro , KIMIZUKA Ryo

    … <p>Open-Domain Question-Answering is the combined task of machine reading comprehension and information retrieval. … To solve this task, we propose 3 approaches. … 1.Query expansion from reading comprehension dataset, 2.Normalize reading comprehension output via sigmoid function, 3. …

    Proceedings of the Annual Conference of JSAI JSAI2020(0), 4Q2GS901-4Q2GS901, 2020

    J-STAGE 

  • Learning to Set Planning Goals for TextWorld by Deep Neural Semantic Parsing  [in Japanese]

    AGRAVANTE Don Joven , TATSUBORI Michiaki

    … <p>Our task is learning to set the goals of a planner to solve text-based adventure games, specifically <u>TextWorld</u>, where each game has a different quest that is described in natural language. … This agent architecture bridges the gap between neural networks and classical planning in a novel way by grounding the Transformer output into the <u>PDDL</u> …

    Proceedings of the Annual Conference of JSAI JSAI2020(0), 3F5ES205-3F5ES205, 2020

    J-STAGE 

  • Task Visualization and Job Organization by "Color": Based on Structured Knowledge  [in Japanese]

    KOBAYAKAWA Maiko , IJUIN Koki , SUNAGA Takeshi , NISHIMURA Takuichi

    … In this paper, we propose a system named "Care dignity" as output of this project and report a process of this design and research.</p> …

    Proceedings of the Annual Conference of JSAI JSAI2020(0), 3C5OS23b01-3C5OS23b01, 2020

    J-STAGE 

  • Policy Transfer in Reinforcement Learning with Domain Adaptation using Transition Probability  [in Japanese]

    SATO Rei , FUKUCHI Kazuto , SAKUMA Jun , AKIMOTO Youhei

    … <p>Reinforcement learning is drawing increasing attentions in real world applications.Since it often takes enormous cost to learn the agent in the real world environment (called target task), pre-training in a low-cost environment such as a simulator (called source task) is gathering attention. …

    Proceedings of the Annual Conference of JSAI JSAI2020(0), 2J5GS203-2J5GS203, 2020

    J-STAGE 

  • Identifying the Corresponding CT Slices among the Different Scans via Deep Metric Learning  [in Japanese]

    ONGGO Barata Tripramudya , OTA Joji , HORIKOSHI Takuro , YOKOTA Hajime , MORI Yasukuni , SUYARI Hiroki

    … In the case studies, we showed that on a task to pinpoint the matching CT slices of different dates, a network trained with the proposed method can discover similar objects between slices and gives an output equivalent to visual assessment by an experienced radiologist.</p> …

    Proceedings of the Annual Conference of JSAI JSAI2020(0), 2H5GS1305-2H5GS1305, 2020

    J-STAGE 

  • The Utility of Applying Various Image Preprocessing Strategies to Reduce the Ambiguity in Deep Learning-based Clinical Image Diagnosis

    Tachibana Yasuhiko , Obata Takayuki , Kershaw Jeff , Sakaki Hironao , Urushihata Takuya , Omatsu Tokuhiko , Kishimoto Riwa , Higashi Tatsuya

    … <p>Purpose: A general problem of machine-learning algorithms based on the convolutional neural network (CNN) technique is that the reason for the output judgement is unclear. … The strategy is to preprocess the input image data in different ways to highlight the most important aspects of the images for reaching the output judgement.</p><p>Materials and Methods: T<sub>2</sub>-weighted brain image series falling into two age-ranges were used. …

    Magnetic Resonance in Medical Sciences 19(2), 92-98, 2020

    J-STAGE 

  • Development of a Real-Time Force and Temperature Sensing System with MEMS-LSI Integrated Tactile Sensors for Next-Generation Robots

    Muroyama Masanori , Hirano Hideki , Shao Chenzhong , Tanaka Shuji

    … For real-time temperature and force data acquisition, a time-sharing force and temperature task processing mechanism was implemented with a dedicated computer architecture in the FPGA configuration and the host program. … With a fabricated sensor node, sensor output data were measured and analyzed with varying temperatures and applied force. …

    Journal of Robotics and Mechatronics 32(2), 323-332, 2020

    J-STAGE 

  • Evaluation of Dialogue Systems with Different End of Sentences Using Seq2Seq Model  [in Japanese]

    NAGAI Nozomu , MIYAMOTO Tomoki , KATAGAMI Daisuke

    … <p>The purpose of this study is to change the impression given to the user by changing the words of the sentence output by the dialogue system to intimate words or words with respect using the politeness theory known as the consideration for constructing the smooth human relation by the words and deeds of each other. … After learning, we modified the output so that the Seq2Seq model uses two kinds of words preferentially. …

    Journal of Japan Society for Fuzzy Theory and Intelligent Informatics 32(1), 523-527, 2020

    J-STAGE 

  • Knowledge Discovery from Layered Neural Networks Based on Non-negative Task Matrix Decomposition

    WATANABE Chihiro , HIRAMATSU Kaoru , KASHINO Kunio

    … Since a trained layered neural network consists of a complex nonlinear relationship between large number of parameters, we failed to understand how they could achieve input-output mappings with a given data set. … In this paper, we propose the non-negative task matrix decomposition method, which applies non-negative matrix factorization to a trained layered neural network. …

    IEICE Transactions on Information and Systems E103.D(2), 390-397, 2020

    J-STAGE 

  • Japanese Legal Term Correction using Random Forest  [in Japanese]

    Yamakoshi Takahiro , Ogawa Yasuhiro , Komamizu Takahiro , Toyama Katsuhiko

    … In this paper, we first define input and output of legal term correction task. …

    Transactions of the Japanese Society for Artificial Intelligence 35(1), H-J53_1-14, 2020

    J-STAGE 

  • Hidden Markov Model-based Extraction of Target Objects in X-ray Image Sequence for Lung Radiation Therapy  [in Japanese]

    Shindo Masahiro , Ichiji Kei , Homma Noriyasu , Zhang Xiaoyong , Okuda Shungo , Sugita Norihiro , Yamaki Shunsuke , Takai Yoshihiro , Yoshizawa Makoto

    … <p>It is an important task to accurately track the target tumor with respiratory movement during radiation therapy. … We first simulate possible combinations of image intensities of target objects as hidden states and observable X-ray image intensities as output symbol in HMM by using digitally reconstructed radiographs generated from four-dimensional X-ray computed tomography. …

    IEEJ Transactions on Electronics, Information and Systems 140(1), 49-60, 2020

    J-STAGE 

  • Conversations as a Holistic Assessment Tool : An Action Research Report

    Branden KIRCHMEYER

    … This paper presents data collected from student performances on an experimental spoken English assessment task conducted at a Japanese university. … Designed as the foundation for a reconstructed English communication curriculum for first-year university students enrolled in compulsory English classes (Tempest, 2018), the assessment task elicited free-form conversation, which was then transcribed and analyzed for specific metrics of spoken L2 output. …

    崇城大学紀要 = Bulletin of Sojo University 45, 41-47, 2020

    IR 

  • Junior high school output activities based on the "Large Grammar" method : Using English as a communication tool.  [in Japanese]

    石田 順 , 金森 玲子 , 竹川 由紀子 , 足立 和美

    … At Tottori University junior high school we practice output activities on a daily basis in our English classes with the aim of enhancing students' ability to communicate in English. … These activities, using the Large Grammar method proposed by one of the authors, Kazumi Adachi, rely on improvised output. … We are aiming for activities that encourage students to think "voluntarily", "interactively" and "deeply" about the task. …

    鳥取大学附属中学校研究紀要 (50), 73-102, 2019-03-01

    IR 

Page Top