Scholars' Association News
Issue 38
May 2016

06/08


Show Images Hide Images

Previous
Next article
Language technology: Language meets modern tech
by ARCHONDOULA MENTI

The massive advances in modern technology over recent years have influenced how people communicate, making communication multimodal, with an increasing presence for visual and audio elements thanks to the use of multimedia. However, the IT revolution has not meant that the linguistic element has been overlooked; quite the opposite, it has showcased language as an effective tool in communications at all levels. Modern technology's achievements demonstrate this point.

Language technology is a scientific discipline where linguistics meets IT and deals with how to integrate language skills into computer apps which are used in ICT systems. Some of language technology's achievements are familiar to us since we use them every day (such as 'smart' keyboards on mobile phones or voice-activated searches on the Google search engine) while others are less well known, because they are part of broader applications (such as filters that read and automatically classify email messages).

Some key sectors where language apps and computing tools are being developed are automatic speech analysis and processing, machine translation, voice synthesis and recognition, and data 'mining' from texts, as well as educational technology. These sectors are based on natural language being processed by computers and are closely tied into human/machine interaction.

In the area of automatic speech processing and analysis we find computer tools designed to separate sentences and words, perform textual morphological analyses, extract lemmas and perform syntax and meaning analyses of the text; tools which are normally used as individual components of software such as machine translation software.

Machine translation means the automated process of translating a language unit (a phrase, sentence, text) from one language to another using a computer program. Modern machine translation software is based on grammatical or statistical processing of language data. Currently there are several machine translation platforms, some of which are freely accessible online, and particularly popular (Google Translate, Bing Translator). As a field, machine translation is currently focusing on achieving higher quality translation and combating the weaknesses and addressing the errors generated by automatic translations, and has also expanded into the machine translation of the spoken word, thanks to the use of automatic speech recognition.

The terms 'voice synthesis' and 'voice recognition' are encountered in the machine production and recognition of human speech. In the first, a text is converted to speech by a text-to-speech (TTS) system while in the second case human speech is converted into text (speech-to-text, STT). Voice synthesis and recognition applications can be found in computers, mobile phones and other electronic devices that have voice-activated keyboards, or allow for voice-activated control of household applications or online searches, and are also built into automatic text generators (for product descriptions, for example) and spoken dialogue systems (such as ticket booking systems or satnavs in cars).

These areas are very important, offering support technologies to the disabled; allowing the visually impaired or those with reading difficulties to 'listen' to written texts and people with mobility difficulties to 'write' anything they are able to say, thereby improving their day-to-day life and making it easier for them to access the information society on an equal footing.

The field of data 'mining' or extraction is based on recognising specialist information within a dataset and automatically retrieving it. It is primarily aimed at automatic, mechanical identification of names (people's names, place names, names of organisations, etc.), chronological data (dates) and events, and relationships between them within digital texts. The results of data mining are used in a wide number of applications such as automatic text summaries, Q&A systems (searches for specific information on websites) and machine translation. Current fields of research interest in this sector include 'mining' opinions and analysis and evaluation of emotions though internet texts (such as blogs and social network posts).

Finally, 'educational technology' from the viewpoint of language technology means the technology that contributes to language education via multimodal educational tools and digital language activities, to improve language acquisition and solve problems in teaching mother tongues or foreign languages. Here methods and tools are being developed (educational planning, audiovisual teaching materials, training materials, educational equipment, diagnostic tools, collection and processing of language resources, creation of online lexicographical tools, etc.) which assist in language teaching, evaluate and address learning difficulties and speech disorders in children and foster learning at your own pace, while also cutting the time it takes to learn a foreign language.

Language technology is a vital element of the digital society. From machine translation to voice synthesis and recognition, its applications and services are improving our day-to-day lives, expanding our knowledge of language and computing mechanisms, and above all offer people the ability to communicate more effectively with their computers, and especially with each other.

(Archondoula Menti is a linguist).


‹ Previous  |  Next article ›