Technology Pipeline for Large Scale Cross-Lingual Dubbing of Lecture Videos into Multiple Indian Languages

Cross-lingual dubbing of lecture videos requires the transcription of the original audio, correction and removal of disfluencies, domain term discovery, text-to-text translation into the target language, chunking of text using target language rhythm, text-to-speech synthesis followed by isochronous lipsyncing to the original video. This task becomes challenging when the source and target languages belong to different language families, resulting in differences in generated audio duration. This is further compounded by the original speaker's rhythm, especially for extempore speech. This paper describes the challenges in regenerating English lecture videos in Indian languages semi-automatically. A prototype is developed for dubbing lectures into 9 Indian languages. A mean-opinion-score (MOS) is obtained for two languages, Hindi and Tamil, on two different courses. The output video is compared with the original video in terms of MOS (1-5) and lip synchronisation with scores of 4.09 and 3.74, respectively. The human effort also reduces by 75%.

Evaluating MT Systems: A Theoretical Framework

This paper outlines a theoretical framework using which different automatic metrics can be designed for evaluation of Machine Translation systems. It introduces the concept of {\em cognitive ease} which depends on {\em adequacy} and {\em lack of fluency}. Thus, cognitive ease becomes the main parameter to be measured rather than comprehensibility. The framework allows the components of cognitive ease to be broken up and computed based on different linguistic levels etc. Independence of dimensions and linearly combining them provides for a highly modular approach. The paper places the existing automatic methods in an overall framework, to understand them better and to improve upon them in future. It can also be used to evaluate the newer types of MT systems, such as speech to speech translation and discourse translation.

LERIL : Collaborative Effort for Creating Lexical Resources

The paper reports on efforts taken to create lexical resources pertaining to Indian languages, using the collaborative model. The lexical resources being developed are: (1) Transfer lexicon and grammar from English to several Indian languages. (2) Dependencey tree bank of annotated corpora for several Indian languages. The dependency trees are based on the Paninian model. (3) Bilingual dictionary of 'core meanings'.

Language Access: An Information Based Approach

The anusaaraka system (a kind of machine translation system) makes text in one Indian language accessible through another Indian language. The machine presents an image of the source text in a language close to the target language. In the image, some constructions of the source language (which do not have equivalents in the target language) spill over to the output. Some special notation is also devised. Anusaarakas have been built from five pairs of languages: Telugu,Kannada, Marathi, Bengali and Punjabi to Hindi. They are available for use through Email servers. Anusaarkas follows the principle of substitutibility and reversibility of strings produced. This implies preservation of information while going from a source language to a target language. For narrow subject areas, specialized modules can be built by putting subject domain knowledge into the system, which produce good quality grammatical output. However, it should be remembered, that such modules will work only in narrow areas, and will sometimes go wrong. In such a situation, anusaaraka output will still remain useful.

Anusaaraka: Overcoming the Language Barrier in India

The anusaaraka system makes text in one Indian language accessible in another Indian language. In the anusaaraka approach, the load is so divided between man and computer that the language load is taken by the machine, and the interpretation of the text is left to the man. The machine presents an image of the source text in a language close to the target language.In the image, some constructions of the source language (which do not have equivalents) spill over to the output. Some special notation is also devised. The user after some training learns to read and understand the output. Because the Indian languages are close, the learning time of the output language is short, and is expected to be around 2 weeks. The output can also be post-edited by a trained user to make it grammatically correct in the target language. Style can also be changed, if necessary. Thus, in this scenario, it can function as a human assisted translation system. Currently, anusaarakas are being built from Telugu, Kannada, Marathi, Bengali and Punjabi to Hindi. They can be built for all Indian languages in the near future. Everybody must pitch in to build such systems connecting all Indian languages, using the free software model.

Information Revolution

The world is passing through a major revolution called the information revolution, in which information and knowledge is becoming available to people in unprecedented amounts wherever and whenever they need it. Those societies which fail to take advantage of the new technology will be left behind, just like in the industrial revolution. The information revolution is based on two major technologies: computers and communication. These technologies have to be delivered in a COST EFFECTIVE manner, and in LANGUAGES accessible to people. One way to deliver them in cost effective manner is to make suitable technology choices (discussed later), and to allow people to access through shared resources. This could be done throuch street corner shops (for computer usage, e-mail etc.), schools, community centers and local library centres.

Collaborative Creation of Digital Content in Indian Languages

The world is passing through a major revolution called the information revolution, in which information and knowledge is becoming available to people in unprecedented amounts wherever and whenever they need it. Those societies which fail to take advantage of the new technology will be left behind, just like in the industrial revolution. The information revolution is based on two major technologies: computers and communication. These technologies have to be delivered in a COST EFFECTIVE manner, and in LANGUAGES accessible to people. One way to deliver them in cost effective manner is to make suitable technology choices, and to allow people to access through shared resources. This could be done throuch street corner shops (for computer usage, e-mail etc.), schools, community centres and local library centres.

Anusaaraka: Machine Translation in Stages

Fully-automatic general-purpose high-quality machine translation systems (FGH-MT) are extremely difficult to build. In fact, there is no system in the world for any pair of languages which qualifies to be called FGH-MT. The reasons are not far to seek. Translation is a creative process which involves interpretation of the given text by the translator. Translation would also vary depending on the audience and the purpose for which it is meant. This would explain the difficulty of building a machine translation system. Since, the machine is not capable of interpreting a general text with sufficient accuracy automatically at present - let alone re-expressing it for a given audience, it fails to perform as FGH-MT. FOOTNOTE{The major difficulty that the machine faces in interpreting a given text is the lack of general world knowledge or common sense knowledge.}

An Algorithm for Aligning Sentences in Bilingual Corpora Using Lexical Information

In this paper we describe an algorithm for aligning sentences with their translations in a bilingual corpus using lexical information of the languages. Existing efficient algorithms ignore word identities and consider only the sentence lengths (Brown, 1991; Gale and Church, 1993). For a sentence in the source language text, the proposed algorithm picks the most likely translation from the target language text using lexical information and certain heuristics. It does not do statistical analysis using sentence lengths. The algorithm is language independent. It also aids in detecting addition and deletion of text in translations. The algorithm gives comparable results with the existing algorithms in most of the cases while it does better in cases where statistical algorithms do not give good results.