Linguistic Indicators for Estimating the Quality of Machine Translations

Add to your list(s) Download to your calendar using vCal

• Mariano Felice, University of Cambridge
• Friday 15 February 2013, 12:00-13:00
• FW26, Computer Laboratory.

If you have a question about this talk, please contact Ekaterina Kochmar.

This talk presents a summary of research done as part of my master’s thesis, where I have addressed the problem of estimating the quality of machine translations automatically without having access to human references. Estimations are obtained by building a supervised regression model that predicts quality scores using features from the source and target texts as well as additional resources. Although shallow indicators are generally used to characterise the relationship between the source text and its translation, they convey no notion of meaning, grammar or linguistic correctness so final estimations may be very biased towards superficial aspects. The work presented here attempts to bridge that gap by introducing more linguistic features and analysing how they compare with shallow indicators over a set of publicly available datasets.

This talk is part of the NLIP Seminar Series series.

This talk is included in these lists:

• All Talks (aka the CURE list)
• Cambridge Centre for Data-Driven Discovery (C2D3)
• Cambridge Forum of Science and Humanities
• Cambridge Language Sciences
• Cambridge talks
• Chris Davis' list
• Department of Computer Science and Technology talks and seminars
• FW26, Computer Laboratory
• Graduate-Seminars
• Guy Emerson's list
• Interested Talks
• Language Sciences for Graduate Students
• NLIP Seminar Series
• PMRFPS's
• Simon Baker's List
• Trust & Technology Initiative - interesting events
• bld31
• ndk22's list
• ob366-ai4er
• rp587

Note that ex-directory lists are not shown.