Summer 2005 V-Unit Projects

Time: Wednesday, October 26, 2005, 4:30 – 7:00 p.m. Venue: Wean Hall 4509 See the poster

• Schedule of Presenters:
• 4:30 – 5:00 p.m. Introduction to the V-Unit, Manuela Veloso, Bernardine Dias, and Juan Pablo Gonzalez
• 5:00 – 5:30 p.m. Ayorkor Mills-Tettey, Ph.D. Student, RI, “Investigating the Impact of an Automated Reading Tutor in Ghana”
• 5:30 – 6:00 p.m. Aditya Chand, M.S. Student, HCII, “Physical Interfaces for Semiliterate Users”
• 6:00 – 6:30 p.m. Fei Huang, Ph.D. Student, LTI, “Detecting Bilingual Medical Term Translations via Web Mining”
• 6:30 – 7:00 p.m. Ariadna Font Llitjós, Ph.D. Student, LTI, “Developing a Quechua to Spanish Translation System”

Investigating the Impact of an Automated Reading Tutor in Ghana, Ayorkor Mills-Tettey

Abstract

One contributing factor to the low reading proficiency of some children in Ghana may be lack of opportunity to practice reading, particularly guided reading. For the average Ghanaian child from a rural or low-income urban background, reading may not be part of daily family life and schools typically offer few opportunities for individual attention while developing reading skills. This field study investigates the effects of two approaches to guided reading practice: using the LISTEN Reading Tutor and regular reading practice with an older, literate youth. The LISTEN (Literacy Innovation that Speech Technology Enables) reading tutor is an automated reading tutor, in development at Carnegie Mellon University since 1996, that uses speech recognition and artificial intelligence to analyze a child’s reading and give graphical and spoken feedback. This initial pilot phase of the study in Ghana aims to investigate the feasibility of the technology in the Ghanaian setting. The second phase will be aimed at measuring the efficacy of the reading tutor compared to regular practice with an older youth, in helping children improve their reading skills.

Bio

Ayorkor Mills-Tettey is in the Ph.D. program in the Robotics Institute at Carnegie Mellon University. She grew up in Nigeria and Ghana, and then moved to the U.S. for university in 1997. She majored in Computer Science at Dartmouth College in New Hampshire, and then stayed at Dartmouth for a Masters Degree at the Thayer School of Engineering. In 2003, she returned to Ghana to spend a year teaching computer science at Ashesi University in Accra before joining the Robotics Institute in 2004. Ayorkor’s current research focuses on path planning for robotics exploration. She is also very interested in understanding the synergies between technology and education, especially in developing communities. She spent time in Ghana in the summer of 2005, conducting a pilot project using an automated reading tutor to assist students with English language literacy.

Physical Interfaces for Semiliterate Users, Aditya Chand

Abstract

In developing countries, like India, Illiteracy, non-native language, Computer illiteracy and fear of technology are big hurdles towards getting people to utilize the power of the digital medium/internet. People do not recognize the power of the medium and therefore do not invest resources (time and money) to overcome the learning curve. This is a big hurdle towards creating the critical mass to boost the use of communication methods like email and online communities. The project aims to help the semi-literate population in rural areas in India to access online information through the use of physical interfaces. The use of physical and visual metaphors can help the novice users overcome the barriers of learning language (English) and technology (mouse and keyboard). This interface would provide its users an initial glimpse of the power of internet which could be a motivation for them to learn more about the language (English and computers).

Bio

Aditya Chand is a full time Masters Student at HCII. He is interested in cross cultural design, information visualization and collaborative environments. Aditya holds a Bachelors in Visual Communication Design from IIT Guwahati, India. Before coming to CMU he worked on Interaction Design projects at Media Lab Asia and tackled Usability issues at Human Factors International, Mumbai. Apart from design, he enjoys traveling, photography and basketball.

Detecting Bilingual Medical Term Translations via Web Mining, Fei Huang

Abstract

Rapid economic growth and globalization enable unprecedented culture exchange, which also presents great challenge to the communication in a multilingual world. Because of language barriers, people speaking different languages find it difficult to communicate efficiently, especially in urgent situations, e.g., an emergency room in a hospital. Therefore medical dialogue translation systems are in great need. In this project Fei will first interview doctors and patients, asking their experience on how language barriers hinder the effective communication between non-English speaking patients and English speaking doctors. Then Fei will survey several medial dialogue translation systems that aim to overcome these challenges. Finally Fei will develop an English-Chinese medical term translation system via web mining, and the developed approach will be potentially applied to other language pairs.

Bio

Fei Huang is a Ph.D. candidate in Language Technologies Institute, School of Computer Science, Carnegie Mellon University. He graduated with a B.S. in Electrical Engineering and B.A. in Linguistics in 1996 from Tianjin University in China and M.Sc. in Pattern Recognition and Intelligent System in 1999 from Chinese Academy of Sciences. In 2001 He completed his M.Sc. in language and information technologies at CMU. His research interests are in machine translation, information extraction, spoken language processing and statistical machine learning.

Developing a Quechua to Spanish Translation System, Ariadna Font Llitjós

Abstract

Information and communication technology (ICT) is playing an increasingly influential role in reshaping trade, growth, employment and production in large parts of the world. ICT presents unprecedented opportunities to combat poverty by increasing income, opening markets and providing a channel through which the voices of the poor can be heard. One of the world’s main challenges is to broaden the reach of ICT to those who risk being left behind. Ariadna’s thesis research, interactive and automatic refinement of translation rules, is part of a project concerned with the rapid and affordable development of MT systems for resource-poor languages, such as Quechua and Mapudungun. The V Unit has allowed Ariadna to spend this past summer in Peru developing the infrastructure for a Quechua to Spanish translation system prototype. Having such a prototype will allow us to test the tools and algorithms that were designed and implemented for other language pairs, for Quechua-Spanish.

Bio

Ariadna Font Llitjós was born in Barcelona, Catalunya (Spain), and graduated from the Universitat Pompeu Fabra (UPF) in 1996 with a bachelor’s degree in Translation and Interpreting (English and German). She then worked at the Institut de Lingüística Aplicada (IULA) at UPF developing a Catalan transfer grammar for a Machine Translation System (LEKTA) and at the Institut für Maschinelle Sprachverarbeitung (IMS) at Universität Stuttgart. In 1997, Ms. Font Llitjós joined a Cognitive Sciences and Language Ph.D. program at UPF, Universitat Autònoma de Barcelona (UAB), Universitat de Barcelona (UB) and Universitat Politècnica de Catalunya (UPC). She was awarded a “La Caixa” fellowship in 1999 to pursue graduate studies in the U.S. She completed her MS in language and information technologies in 2001, and is currently pursuing her Ph.D. at the Language Technologies Institute (LTI) at Carnegie Mellon University. Ariadna’s current research centers on developing Automatic Machine Translation Systems for resource-poor languages, such as Mapudungun and Quechua.