Extending feature decay algorithms using
alignment entropy
Poncelas, AlbertoORCID: 0000-0002-5089-1687, Toral, AntonioORCID: 0000-0003-2357-2960 and Way, AndyORCID: 0000-0001-5736-5930
(2017)
Extending feature decay algorithms using
alignment entropy.
In: FETLT 2016: Future and Emerging TrenFETLT 2016: Future and Emerging Trends in Language Technologies, Machine Learning and Big Datauage Technologies, Machine Learning and Big Data. 2nd International Workshop, 30 Nov- 2 Dec 2016, Seville, Spain.
In machine-learning applications, data selection is of crucial importance if good runtime performance is to be achieved. Feature Decay Algorithms (FDA) have demonstrated excellent performance in a number of
tasks. While the decay function is at the heart of the success of FDA,
its parameters are initialised with the same weights. In this paper, we
investigate the effect on Machine Translation of assigning more appropriate weights to words using word-alignment entropy. In experiments on
German to English, we show the effect of calculating these weights using two popular alignment methods, GIZA++ and FastAlign, using both
automatic and human evaluations. We demonstrate that our novel FDA
model is a promising research direction.
Metadata
Item Type:
Conference or Workshop Item (Paper)
Event Type:
Conference
Refereed:
Yes
Uncontrolled Keywords:
Data selection; Machine translation; Mathematical foundations
Proceedings of FETLT 2016: Future and Emerging Trends in Language Technologies, Machine Learning and Big Data. Lecture Notes in Computer Science
10341.
Springer.
This item is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 3.0 License. View License
Funders:
ADAPT Centre for Digital Content Technology, funded under the SFI Research Centres Programme (Grant 13/RC/2106), European Regional Development Fund, and the European Union Seventh Framework Programme FP7/2007-2013 under grant agreement PIAP-GA-2012-324414 (AbuMaTran)
ID Code:
23232
Deposited On:
02 May 2019 12:00 by
Thomas Murtagh
. Last Modified 22 Jan 2021 14:17