Modelling Annotator Bias with Multi-task Gaussian Processes: An Application to Machine Translation Quality Estimation

By Trevor Cohn, Lucia Specia. 2013

123 citations recorded by the source corpus.

Subfields: Probabilistic & Bayesian Methods, Machine Translation

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Probabilistic & Bayesian Methods

In Machine Translation

Citations in the corpus

Non-Linear Regression for Bag-of-Words Data via Gaussian Process Latent Variable Set Model (2015, background): “All rights reserved. such regression problems in various domains, e.g. natural language processing (Cohn and Specia 2013), time series analysis (Preotiuc-Pietro and Cohn 2013), computer vision (Kapoor et al. 2009) and data mining (Lampos and Aletras 2014).”
SHEF-Lite: When Less is More for Translation Quality Estimation (2013, methodology): “While a variety of kernel functions are available, here we followed previous work on QE using GP (Cohn and Specia, 2013; Shah et al., 2013) and employed a squared exponential (SE) kernel with automatic relevance determination (ARD): where F is the number of features, σ 2 f is the co-variance…”
Findings of the 2013 Workshop on Statistical Machine Translation (2013, background): “Comparison with the better results seen on Tasks 1.1 and 1.2, however, suggests that, for Task 1.3, additional data processing, filtering, and modelling (including modelling translator-specific traits such as their variance in time) is required, as evidenced in (Cohn and Specia, 2013).”
Predicting and Characterising User Impact on Twitter (2014, background): “Recently, it was shown by Cohn and Specia (2013) that Support Vector Machines for Regression (SVR) (Vap-nik, 1998; Smola and Sch¨olkopf, 2004), commonly considered the state-of-the-art for NLP regression tasks, can be outperformed by Gaussian Processes (GPs), a kernelised, probabilistic approach to…”
A temporal model of text periodicities using Gaussian Processes (2013, methodology): “(2011) and for Quality Estimation regression in (Cohn and Specia, 2013; Shah et al., 2013).”
Strong Baselines for Neural Semi-Supervised Learning under Domain Shift (2018, background): “Recent NLP conferences witnessed a “tsunami” of deep learning papers (Manning, 2015), followed by what we call a multi-task learning “wave”: MTL has been successfully applied to a wide range of NLP tasks (Cohn and Specia, 2013; Cheng et al., 2015; Luong et al., 2015; Plank et al., 2016; Fang and Coh”
SHEF-Lite 2.0: Sparse Multi-task Gaussian Processes for Translation Quality Estimation (2014, methodology): “While a variety of kernel functions are available, here we followed previous work in QE using GP (Cohn and Specia, 2013; Shah et al., 2013) and employed a squared exponential (SE) kernel with automatic relevance determination (ARD):”
Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotations (2021, background): “Cohn and Specia (2013) use a multi-task approach to model annotator differences in machine translation annotations.”
Online Multitask Learning for Machine Translation Quality Estimation (2015, background): “On the other way round, task relatedness is successfully exploited by Cohn and Specia (2013), who apply multitask learning to jointly learn from data obtained from several annotators with different levels of expertise and reliability.”