DOCSLIB.ORG

Voice Typing”, Which Reduces the Recognition Error Rate in Modern Day Dictation Interfaces Simply by Modifying the User’S Interaction Style

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Voice Typing”, Which Reduces the Recognition Error Rate in Modern Day Dictation Interfaces Simply by Modifying the User’S Interaction Style Enabling Non-Speech Experts to Develop Usable Speech-User Interfaces Thesis Proposal Anuj Kumar [email protected] Human-Computer Interaction Institute School of Computer Science Carnegie Mellon University Thesis Committee: Florian Metze, CMU (chair) Matthew Kam, CMU & American Institutes for Research (co-chair) Dan Siewiorek, CMU Tim Paek, Microsoft Research Abstract Speech user interfaces (SUIs) such as Apple’s Siri, Samsung’s S Voice, and Google’s Voice Search are becoming increasingly popular. However, despite years of research, speech interfaces are only actively being developed for specific users, such as adult native speakers of English, when in fact, less mainstream users such as non-native English speakers or children stand to benefit at least as much, if not more. The problem in developing SUIs for such users or for challenging acoustic conditions is the expertise, time, and cost in building a reasonably accurate speech recognition system. In particular, researchers who are not speech experts find it difficult to build a working speech interface on their own, and instead routinely resort to Wizard-of-Oz experiments for testing their application ideas. Unfortunately, such experiments are unable to test real usage of the system. At the same time, speech recognition technology has reached a point where, given enough expertise and in-domain data, a working recognizer can, in principle, be developed for almost every user or acoustic condition. To address the above problem, in this thesis, I propose to design, develop, and evaluate a toolkit that lowers the threshold for developing accurate speech systems. In particular, my research goals are three-fold: (i) to discover and formalize the tacit knowledge that speech experts employ while optimizing a recognizer for greater accuracy, (ii) to investigate approaches for supporting the semi-automatic analysis of errors occurring in speech recognition, and (iii) to investigate strategies for improving recognition accuracy. Researchers who are non-specialists in speech recognition can benefit from the proposed toolkit because it will assist them in developing speech recognizers with less cost, time, and expertise, which they would otherwise not be likely to develop and experiment with. The goal is not to render the speech expert superfluous, but to make it easier for non-speech experts, i.e. the developers utilizing speech technology, to figure out why a speech system is failing, and guide their efforts in the right direction. I have accomplished several steps towards the above goals. Using contextual interviews and cognitive task analysis, I have formalized the intuition and technical know-how that speech experts draw on when improving recognition accuracy, in the form of a rule-based knowledge base. I have built a toolkit that supports semi-automatic error analysis, and embeds this rule-based knowledge base. Furthermore, by applying these rules to two datasets, I have demonstrated that the recommendations from the knowledge base can lead to the development of recognizers with accuracy at least as much as those developed using experts’ recommendations. Once recognizers have been reasonably optimized for the target users, we make the case that subtle changes to the user interaction can lead to further gains in usability. To demonstrate the above, I discuss the design and evaluation of a novel technique called “Voice Typing”, which reduces the recognition error rate in modern day dictation interfaces simply by modifying the user’s interaction style. We see such techniques as a part of the strategies that the toolkit will eventually recommend. For the remainder of my PhD, I propose two experiments: first, to evaluate the extent to which non-speech experts can benefit from the guidance provided by the toolkit, and second, to measure the usability gains that the accuracy improvements translates into for the end-users. We will test both the above experiments in the context of a children’s dataset that has over 100 hours of spontaneous and scripted speech. This dataset will enable us to test the efficacy of the toolkit on a challenging speech problem, i.e. large-vocabulary continuous speech recognition for children. 2 The proposed work attempts to take the “black art” out of speech-user interface development. It contributes to human-computer interaction by developing rapid-development tools for non-speech experts to develop usable SUIs. It contributes to speech technologies by formalizing speech experts’ knowledge and offering a set of tools to analyze speech data systematically. 3 Table of Contents 1. Introduction 1.1. Current Approaches 1.2. Proposed Approach 1.3. Research Goals 1.4. Contributions of this Thesis 1.5. Thesis Proposal Organization 2. Case Study #1: Speech Games for Second Language Learners 2.1. Need for Speech-enabled Second Language Learning Games 2.2. English Language Competency: Word Reading 2.3. Game Designs 2.4. Initial Usability Testing 2.5. Speech Recognizer Development and Improvements 2.6. Subsequent User Study 2.7. Discussion 3. Case Study #2: Voice Typing – A Novel Interaction Technique 3.1. Voice Typing 3.2. Related Work: Real-time Speech Decoding and Error Correction Techniques 3.3. User Study 3.4. Results 3.5. Discussion 4. Speech Interfaces: Myths vs. Empirical Reality 4.1. Myth #1: No data like more data 4.2. Myth #2: SUIs are unusable by non-native speakers, children, or in noisy environments 4.3. Myth #3: SUIs need to achieve very high accuracy before they are usable 4.4. Myth #4: SUIs are complex and expensive to build 5. Capturing and Representing Experts’ Understanding in a Knowledge Base 5.1. Knowledge Elicitation Methodology: Contextual Interviews with Speech Experts 5.2. Related Work: Expert Systems and Rule-based Knowledge Representation 5.3. Primer to Speech Recognition 5.4. Knowledge Formalization 5.4.1. ChAOTIC Process Framework and Rules 5.4.2. Characteristics of Knowledge Base 5.4.3. Expert System and Forward Chaining Inference Engine 5.4.4. Conflict Resolution 5.5. Datasets 5.6. Data Annotation 5.7. Evaluation Experiments 5.8. Discussion and Next Steps 4 6. Design and Development of SToNE: Speech Toolkit for Non-Experts 6.1. Related Work: Error Analysis in Speech Recognition 6.1.1. Types of Errors 6.1.2. Lexical and Prosodic Reasons of Error 6.2. SToNE Design and Development 6.2.1. Feature Extractor 6.2.2. Error Visualizer: Data Selector & Performance Comparer 6.2.3. Optimization Advisor 7. Proposed Work 7.1. Evaluation of Toolkit with Non-Speech Experts 7.2. Evaluation of a Speech Application with End-Users 7.3. Risks and Fallback Strategies 8. Thesis Overview and Timeline 9. References 5 1. Introduction Speech-user interfaces (SUIs) are becoming increasingly popular. For users in emerging regions, where illiteracy often impedes usage of graphical and textual interfaces, spoken dialog systems are being explored in many domains, such as agriculture [Plauche, ’06], health [Sherwani, ’09], entertainment service [Raza, ’13], etc. Similarly, for users in the developed world, several speech interfaces have gained traction. United Airlines’ customer care system, which receives an average of two million calls per month, indicated that their caller abandonment dropped by 50% after implementing a speech system to replace a touchtone system [Kotelly, ’03]. Such a solution introduced shallower menu hierarchies, eyes-free, hands-free operation, and ease of data entry for non-numeric text [Hura, ’08]. Recently introduced mobile assistants like Apple’s Siri, Samsung’s S Voice, and Google’s Skyvi have further boosted the popularity of speech applications. They assist in executing simple commands such as “Call John”, “Text my wife that I’ll be late today”, or “Set a wakeup alarm for tomorrow 6am!” to more complicated tasks such as dictation or transcribing a phone conversation. These examples underscore the reasons why many researchers and practitioners are becoming increasingly interested in incorporating speech as an input modality in their applications. Yet, it is telling that beyond a few examples, we haven’t seen a large number of functional and sufficiently accurate speech services. The reason is the difficulty in building a usable speech user interface (SUI) for a certain user group or acoustic situation, without having a good, off-the-shelf speech recognizer available [Laput, ’13]: recognizers work reliably when it is known what people say, how they say it, what noises do occur, etc. This is hard to achieve, especially for SUI designers who are non-experts in speech recognition, such as designers in human-computer interaction. As a result, many studies that HCI researchers have conducted are “Wizard-of-Oz” studies, i.e. using simulated ASR [e.g., Stifelman et al., ’13; Tewari, ’13]. Unfortunately, while Wizard-of-Oz studies are good for iterating on the design of the application, they cannot uncover real recognition errors and test real system usage [Thomason, ’13]. These therefore become the leading source of usability issues in speech applications once the product is released [Lai, Karat & Yankelovich, ’08]. At the same time, research on automatic speech recognition (ASR) has seen great strides in recent times. It has reached a point where given enough speech expertise and in-domain data, a sufficiently accurate speech recognizer can be developed for any scenario, including those of non- native accents, background noise, children’s voice, or other similar challenges for speech recognition. However, if a speech recognition system doesn’t work as intended, it is generally impossible for a non-expert or a HCI researcher to tell, for example, if the ASR fails because users speak too slow (“speaking rate” is a frequent cause of mismatch), too “clearly” (i.e. they “hyper-articulate”), because the background noises are unexpected, or because of some other reason; while an expert can usually analyze the error pattern quickly. “Adaptation” or several optimizations of an existing recognizer can generally mitigate the problem, and will often result in a functional system [Abras & Krimchar, 2004; Preece & Rogers, 2002], but recognizing what to do and how do it needs tremendous expertise and experience in developing speech recognition systems.
Load more

Recommended publications
  • Enabling Non-Speech Experts to Develop Usable Speech-User Interfaces
    Enabling Non-Speech Experts to Develop Usable Speech-User Interfaces Anuj Kumar CMU-HCII-14-105 August 2014 Human-Computer Interaction Institute School of Computer Science Carnegie Mellon University Pittsburgh, PA 15213 USA Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Copyright © Anuj Kumar 2014. All rights reserved. Committee: Florian Metze Co-Chair, Carnegie Mellon University Matthew Kam Co-Chair, Carnegie Mellon University & American Institutes for Research Dan Siewiorek Carnegie Mellon University Tim Paek Microsoft Research The research described in this dissertation was supported by the National Science Foundation under grants IIS- 1247368 and CNS-1205589, Siebel Scholar Fellowship 2014, Carnegie Mellon’s Human-Computer Interaction Institute, and Nokia. Any findings, conclusions, or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the above organizations or corporations. KEYWORDS Human-Computer Interaction, Machine Learning, Non-Experts, Rapid Prototyping, Speech-User Interfaces, Speech Recognition, Toolkit Development. II Dedicated to Mom and Dad for their eternal support, motivation, and love III IV ABSTRACT Speech user interfaces (SUIs) such as Apple’s Siri, Microsoft’s Cortana, and Google Now are becoming increasingly popular. However, despite years of research, such interfaces really only work for specific users, such as adult native speakers of English, when in fact, many other users such as non-native speakers or children stand to benefit at least as much, if not more. The problem in developing SUIs for such users or for other acoustic or language situations is the expertise, time, and cost in building an initial system that works reasonably well, and can be deployed to collect more data or also to establish a group of loyal users.
    [Show full text]
  • Program Details
    Home Program Hotel Be an Exhibitor Be a Sponsor Review Committee Press Room Past Events Contact Us Program Details Monday, November 3, 2014 08:30-10:00 MORNING TUTORIALS Track 1: An Introduction to Writing Systems & Unicode Presenter: This tutorial will provide you with a good understanding of the many unique characteristics of non-Latin Richard Ishida writing systems, and illustrate the problems involved in implementing such scripts in products. It does not Internationalization provide detailed coding advice, but does provide the essential background information you need to Activity Lead, W3C understand the fundamental issues related to Unicode deployment, across a wide range of scripts. It has proved to be an excellent orientation for newcomers to the conference, providing the background needed to assist understanding of the other talks! The tutorial goes beyond encoding issues to discuss characteristics related to input of ideographs, combining characters, context-dependent shape variation, text direction, vowel signs, ligatures, punctuation, wrapping and editing, font issues, sorting and indexing, keyboards, and more. The concepts are introduced through the use of examples from Chinese, Japanese, Korean, Arabic, Hebrew, Thai, Hindi/Tamil, Russian and Greek. While the tutorial is perfectly accessible to beginners, it has also attracted very good reviews from people at an intermediate and advanced level, due to the breadth of scripts discussed. No prior knowledge is needed. Presenters: Track 2: Localization Workshop Daniel Goldschmidt Two highly experienced industry experts will illuminate the basics of localization for session participants Sr. International over the course of three one-hour blocks. This instruction is particularly oriented to participants who are Program Manager, new to localization.
    [Show full text]
  • Paper, We Describe Three User Studies Undertaken to Investigate Users ’Typing Abilities
    Int. J. Human-Computer Studies 106 (2017) 44–62 Contents lists available at ScienceDirect International Journal of Human-Computer Studies journal homepage: www.elsevier.com/locate/ijhcs Is it too small?: Investigating the performances and preferences of users when typing on tiny QWERTY keyboards Xin Yi a,b,c, Chun Yu a,b,c,∗, Weinan Shi a,b,c, Yuanchun Shi a,b,c a Department of Computer Science and Technology, Tsinghua University, China b Tsinghua National Laboratory for Information Science and Technology, China c Beijing Key Lab of Networked Multimedia, China a r t i c l e i n f o a b s t r a c t Keywords: Typing on tiny QWERTY keyboards on smartwatches is considered challenging or even impractical due to the Text entry limited screen space. In this paper, we describe three user studies undertaken to investigate users ’typing abilities Tiny keyboard and preferences on tiny QWERTY keyboards. The first two studies, using a smartphone as a substitute for a Smartwatch smartwatch, tested five different keyboard sizes (2, 2.5, 3, 3.5 and 4 cm). Study 1 collected typing data from QWERTY participants using keyboards and given asterisk feedback. We analyzed both the distribution of touch points Typing pattern (e.g., the systematic offset and shape of the distribution) and the effect of keyboard size. Study 2 adopted a Bayesian algorithm based on a touch model derived from Study 1 and a unigram word language model to perform input prediction. We found that on the smart keyboard, participants could type between 26.8 and 33.6 words per minute (WPM) across the five keyboard sizes with an uncorrected character error rate ranging from 0.4% to 1.9%.
    [Show full text]
  • | Hao Wanata Utilin at Na Malo Mlet
    |HAO WANATA UTILINUS010019435B2 AT NA MALO MLET (12 ) United States Patent ( 10 ) Patent No. : US 10 ,019 ,435 B2 Ouyang et al. ( 45 ) Date of Patent : * Jul. 10 , 2018 ( 54 ) SPACE PREDICTION FOR TEXT INPUT (56 ) References Cited ( 71 ) Applicant : Google Inc. , Mountain View , CA (US ) U . S . PATENT DOCUMENTS 4 ,534 , 261 A 8 / 1985 Fabrizio ( 72 ) Inventors : Yu Ouyang , San Jose , CA (US ) ; 4 ,833 ,610 A 5 / 1989 Zamora et al. Shumin Zhai , Los Altos, CA (US ) ( Continued ) (73 ) Assignee : Google LLC , Mountain View , CA (US ) FOREIGN PATENT DOCUMENTS ( * ) Notice : Subject to any disclaimer, the term of this CN 1133996 A 10 / 1996 patent is extended or adjusted under 35 CN 1761989 A 4 /2006 U . S . C . 154 ( b ) by 1010 days . (Continued ) This patent is subject to a terminal dis claimer . OTHER PUBLICATIONS Notice of Allowance from U . S . Appl. No . 14 / 196 ,552 , dated Jun . 18 , (21 ) Appl. No. : 14 /452 ,035 2015 , 5 pp . (22 ) Filed : Aug . 5 , 2014 (Continued ) Primary Examiner — Steven Sax (65 ) Prior Publication Data (74 ) Attorney, Agent, or Firm — Shumaker & Sieffert, US 2015 / 0026628 A1 Jan . 22 , 2015 P . A . (57 ) ABSTRACT An example method includes displaying, at a presence Related U . S . Application Data sensitive device , a virtual keyboard , receiving an indication (63 ) Continuation of application No . 13/ 657 , 574 , filed on of input entered at a presence -sensitive device indicating a Oct . 22 , 2012 , now Pat. No . 8 , 819 ,574 . plurality of characters , determining a sequence of at least three character strings in the plurality such that the sequence ( 51 ) Int.
    [Show full text]
  • WORKING PAPERS in LITERACY, CULTURE, and LANGUAGE EDUCATION (WPLCLE) VOLUME 3, April 2014
    WORKING PAPERS IN LITERACY, CULTURE, AND LANGUAGE EDUCATION (WPLCLE) VOLUME 3, April 2014 Department of Literacy, Culture, and Language Education School of Education Indiana University Bloomington EDITORIAL BOARD Founder & Editor‐in‐Chief Serafín M. Coronel‐Molina Managing Editor Leslie Rowland Assistants to the Editors Beth Buchholz Hsiao‐Chin Kuo Alfreda Clegg Erin Lemrow Ying‐Sin Chen Jaehan Park Linda Coggin Stacy Penalva Vesna Dimitrieska Julie Rust Hsiao‐Chun Huang Christy Wessel‐Powell Retno Hendryanti Jae‐Seok Yang Arnell Hammond Pei‐Shan Yu Christina Ivanova Amber Warren Tolga Kargin Bita H. Zakeri Advisory Board Donna Sayers Adomat Larry Mikulecky Stephanie Carter Martha Nyikos James Damico Faridah Pawan D. Ted Hall Beth Lewis Samuelson Mary Beth Hines Raymond Smith Mitzi Lewison Karen Wohlwend Carmen Medina Website Administrators Rebecca Barrett Serafín M. Coronel‐Molina Copyright © 2014 Working Papers in Literacy, Culture, and Language Education (WPLCLE), and the respective authors. All rights reserved. No part of this work may be reproduced in any form by any means, including photocopying and recording, or by any information storage or retrieval system (except for brief quotations in critical articles or reviews) without written permission from WPLCLE or the respective authors. Working Papers in Literacy, Culture, and Language Education (WPLCLE) School of Education, Indiana University W.W. Wright Education Building 201 N. Rose Ave., Room #3044 Bloomington, IN 47405‐1006 Phone: (812) 856‐8270 Fax: (812) 856‐8287 E‐mail: [email protected]
    [Show full text]
  • Contributions of New Technologies to the Teaching of English Pronunciation
    Language Value July 2017, Volume 9, Number 1 pp. 1-35 http://www.e-revistes.uji.es/languagevalue Copyright © 2017, ISSN 1989-7103 Contributions of new technologies to the teaching of English pronunciation Yolanda Joy Calvo Benzies [email protected] University of the Balearic Islands, Spain ABSTRACT One of the most significant changes in language classrooms over recent decades has been the introduction of ICTs. Despite a broad range of previous research in the field, little work has been done to date on assessing the benefits of teaching pronunciation through ICTs, something surprising in view of the large number of existing programs and other materials specifically designed to improve learners’ pronunciation. This paper is intended to contribute to the field in that it will provide an overview of the materials currently available for teaching pronunciation through the use of ICTs, as well as an empirical preliminary study on ESP students’ first contact with using ICTs for learning pronunciation. Results indicate that these students enjoyed using these technological tools and would like to use them again to practise their pronunciation. Keywords: Pronunciation, ESP student’s opinions, apps, software, blogs, websites I. INTRODUCTION Previous research in the teaching and learning of English has shown that more attention is usually paid to written skills than to spoken ones in EFL settings (Alonso 2014, Hornero et al. 2013, Calvo 2016a). Furthermore, within spoken skills, pronunciation has traditionally been neglected, to the point where it has been referred to as the poor relation of the English teaching world (Hughes 2002), the orphan (Gilbert 2010) or even the Cinderella (Underhill 2013) in language lessons.
    [Show full text]
  • Gathering Text Entry Metrics on Android Devices
    Proceedings of the International Conference on Multimedia and Human Computer Interaction Toronto, Ontario, Canada, July 18-19 2013 Paper No. 120 Gathering Text Entry Metrics on Android Devices Steven J. Castellucci, I. Scott MacKenzie York University, Department of Computer Science and Engineering 4700 Keele St., Toronto, Ontario, Canada M3J 1P3 [email protected]; [email protected] Abstract -We developed TEMA, an application to gather Text Entry speed and accuracy Metrics on Android devices. This paper details the features of the application, describes a user study to demonstrate its utility, and establishes entry speed and accuracy measurements for the evaluated text entry techniques. We evaluated and compared four mobile text entry methods: two-thumb QWERTY typing, one-finger QWERTY typing, handwriting recognition, and shape writing recognition. The two QWERTY techniques were the fastest, with no statistically significant difference between them in entry speed or accuracy. Shape writing was slightly slower, but similar in accuracy. Handwriting was the slowest and least accurate technique. Keywords: Text entry, metrics, entry speed, accuracy, Android 1. Introduction Mobile devices are often used for SMS text messaging and social networking. An estimated 9.2 trillion text messages will be sent in 2013 (Web-1) and more than 680 million people access Facebook using mobile devices per month (Web-2). Thus, investigating methods for mobile text entry is a significant research topic. To aid evaluation of mobile text entry methods, we created an application to gather user performance metrics on Android devices: Text Entry Metrics on Android (TEMA). In this paper, we present our motivation for TEMA and its features.
    [Show full text]
  • Iphone Dictation to Text App
    Iphone Dictation To Text App Which Erhart launch so quarterly that Tad poses her peridot? If first-rate or astringent Russ usually curtsies his railroads actuating conceitothergates any or retranslations. fornicating insurmountably and pendently, how chocker is Rustie? Sky remains Mycenaean after Vinod spines dirtily or Honestly, please list your spam folder. You earn also building long texts with flame one tap. Content Philosophy, and types those words as texts in there Text Message field. Convenient editing benefit that provides a mount of suggested words. There and many ways to aggravate and adapt text while dictating Crucially you can. Puts quotation marks around words or phrases. Indeed, coding, you between clip unwanted parts. Be side to read establish your leather after dictation. Call Nicole on her mobile phone. Be trying out there is required to explore the audio to dictation app works on track if you can make no help. Try might be nice for people, cancer can copy your converting text and paste it your different social networks like Facebook, type speech into any search box and incline Start Speech Recognition. It contains numerous languages so you can someone text memos and emails in distinct languages. On the minus side, Notes, regardless of bleach you are? They waste in particular nonetheless the text zones where they can several different projects simultaneously. If spouse are using a mobile device or smartphone, but persist long Pro Versions are not. Siri gets what this mean. Google Voice Typing is a completely free speech to simulate software. Tap the Rewind button commercial skip back.
    [Show full text]
  • Environmental Health Criteria 242
    INTERNATIONAL PROGRAMME ON CHEMICAL SAFETY Environmental Health Criteria 242 DERMAL EXPOSURE IOMC INTER-ORGANIZATION PROGRAMME FOR THE SOUND MANAGEMENT OF CHEMICALS A cooperative agreement among FAO, ILO, UNDP, UNEP, UNIDO, UNITAR, WHO, World Bank and OECD This report contains the collective views of an international group of experts and does not necessarily represent the decisions or the stated policy of the World Health Organization The International Programme on Chemical Safety (IPCS) was established in 1980. The overall objec- tives of the IPCS are to establish the scientific basis for assessment of the risk to human health and the environment from exposure to chemicals, through international peer review processes, as a prerequi- site for the promotion of chemical safety, and to provide technical assistance in strengthening national capacities for the sound management of chemicals. This publication was developed in the IOMC context. The contents do not necessarily reflect the views or stated policies of individual IOMC Participating Organizations. The Inter-Organization Programme for the Sound Management of Chemicals (IOMC) was established in 1995 following recommendations made by the 1992 UN Conference on Environment and Development to strengthen cooperation and increase international coordination in the field of chemical safety. The Participating Organizations are: FAO, ILO, UNDP, UNEP, UNIDO, UNITAR, WHO, World Bank and OECD. The purpose of the IOMC is to promote coordination of the policies and activities pursued by the Participating Organizations, jointly or separately, to achieve the sound management of chemicals in relation to human health and the environment. WHO Library Cataloguing-in-Publication Data Dermal exposure. (Environmental health criteria ; 242) 1.Hazardous Substances - poisoning.
    [Show full text]
  • International Journal of Engineering Sciences
    [Shegdar, 4(7): July, 2015] ISSN: 2277-9655 (I2OR), Publication Impact Factor: 3.785 IJESRT INTERNATIONAL JOURNAL OF ENGINEERING SCIENCES & RESEARCH TECHNOLOGY INTELLIGENT KEYBOARD: A METHOD FOR TEXT INPUT USING TOUCH SCREEN FOR HINDI DEVANAGARI SCRIPT Ms.Supriya A. Shegdar*, Mr.Z. M.Shaikh * Computer Science and Engineering, NK Orchid College of Engineering &Technology,Solapur(Maharastra), India ABSTRACT Now a days, most of the smart phones having touch screen instead of traditional keypad. The virtual keyboard is replaced with traditional keypad. There are various such keyboards that are present in the market including Devanagari language. Devanagari is a script used by many Indian languages like Hindi, Marathi, Nepali and Sanskrit, but typing on such touchscreen is a little difficult for new users. The problems with touchscreen are: - they are highly responsive, Touch area is precise, precision is required for typing and users are experiencing many errors in typing keys with their thumb so as a result, typing on such screen is very slow. The solution to these problems is to use an Intelligent keyboard. An Intelligent keyboard is a virtual keyboard used for touch screen smart phones and tablets where the user enters words by click on one or two letters of the word. The overall objective of this paper is to provide knowledge that can help to improve user experience of Devanagari text entry with an Intelligent keyboard in touch screen mobile devices. KEYWORDS: Smartphone, touch screen, virtual keyboard, Devanagari Script. INTRODUCTION A touchscreen is an input device normally layered on the top of an electronic visual display of an information processing system.
    [Show full text]
  • Comparison of Mobile Operating Systems
    Not logged in Talk Contributions Create account Log in Article Talk Read Edit View history Search Comparison of mobile operating systems From Wikipedia, the free encyclopedia Main page This is a comparison of mobile operating systems (OS). Only the latest versions are shown in the table below, even though older versions may Contents still be marketed. Featured content Current events Contents [hide] Random article 1 About OS Donate to Wikipedia 2 Advanced controls Wikipedia store 3 Application ecosystem Interaction 4 Basic features Help 5 Browser About Wikipedia 6 Communication and connectivity Community portal 7 Language and inputs Recent changes 8 Maps and navigation Contact page 9 Media playback and controls Tools 10 Peripheral support What links here 11 Photo and video Related changes 12 Productivity Upload file 13 Ringtones and alerts Special pages 14 Security and privacy Permanent link Page information 15 Sound and voice Wikidata item 16 Other features Cite this page 17 See also 18 References Print/export Create a book Download as PDF Printable version About OS [ edit ] Languages Legend 中文 Edit links Yes In later versions Beta Limited Tablet-only 3rd party No open in browser PRO version Are you a developer? Try out the HTML to PDF API pdfcrowd.com About OS Ubuntu Feature iOS Android Firefox OS Windows Phone BlackBerry 10 Tizen Sailfish OS Touch Linux Sailfish Canonical Foundation, Alliance, Mer, Ltd. and Open Handset Mozilla Tizen Jolla and Company Apple Inc. Microsoft BlackBerry Ltd. Ubuntu Alliance Foundation Association, Sailfish
    [Show full text]
  • Weighted Finite-State Methods for Spell-Checking and Correction Tommi a Pirinen
    Department of Modern Languages 2014 Weighted Finite-State Methods for Spell-Checking and Correction Tommi A Pirinen Academic dissertation to be publicly discussed, by due permission of the Fac- ulty of Arts at the University of Helsinki in auditorium XII (in lecture room PIII), on the 28th of February at 12 o’clock. University of Helsinki Finland Supervisor Krister Lindén, Helsingin yliopisto, Finland Pre-examiners Prof. Mikel L. Forcada, Universitat d’Alacant, Spain Prof. Lars Borin, Göteborgs universitet, Sweden Opponent Prof. Lars Borin, Göteborgs universitet, Sweden Custos Prof. Lauri Carlson, Helsingin yliopisto, Finland Contact information Department of Modern Languages P.O. Box 24 (Unioninkatu 40) FI-00014 University of Helsinki Finland Email address: [email protected] URL: http://www.helsinki.fi/ Telephone: +358 9 1911, telefax: +358 9 191 51120 Copyright © 2014 Tommi A Pirinen http://creativecommons.org/licenses/by-nc-nd/3.0/deed ISBN 978-952-10-9694-5 (printed) ISBN 978-952-10-9695-2 (PDF) Computing Reviews (1998) Classification: F.1.1, I.2.7, I.7.1 Helsinki 2014 Picaset Oy Weighted Finite-State Methods for Spell-Checking and Correction Tommi A Pirinen Department of Modern Languages P.O. Box 24, FI-00014 University of Helsinki, Finland [email protected] http://www.helsinki.fi/%7etapirine PhD Thesis, Series of Publications, 2014 Helsinki, January 2014, 94 pages ISBN 978-952-10-9694-5 (printed) ISBN 978-952-10-9695-2 (PDF) Abstract This dissertation is a large-scale study of spell-checking and correction using finite-state technology. Finite-state spell-checking is a key method for handling morphologically complex languages in a computationally efficient manner.
    [Show full text]