We present a novel deep learning architecture to address the natural language inference (NLI) task. Existing approaches mostly rely on simple reading mechanisms for independent encoding of the premise and hypothesis. Instead, we propose a novel dependent reading bidirectional LSTM network (DR-BiLSTM) to efficiently model the relationship between a premise and a hypothesis during encoding and in...

Our efforts in the 2013 NLI shared task focused on the potential benefits of external corpora. We show that including training data from multiple corpora is highly effective at robust, cross-corpus NLI (i.e. open-training task 1), particularly when some form of domain adaptation is also applied. This method can also be used to boost performance even when training data from the same corpus is av...

This paper presents an ensemble system combining the output of multiple SVM classifiers to native language identification (NLI). The system was submitted to the NLI Shared Task 2017 fusion track which featured students essays and spoken responses in form of audio transcriptions and iVectors by non-native English speakers of eleven native languages. Our system competed in the challenge under the...

Neuroimaging studies of audiovisual speech processing have exclusively addressed listeners' native language (L1). Yet, several behavioural studies now show that AV processing plays an important role in non-native (L2) speech perception. The current fMRI study measured brain activity during auditory, visual, audiovisual congruent and audiovisual incongruent utterances in L1 and L2. BOLD response...

Digital divide can be considered a macro economical index representing the social differences and the separation between the North and the South of the world. Since the first definition of digital divide, it has been shown that it is also a great and unrecognized problem in the developed countries, especially in the field of education. “Digital disconnection” is a key problem for School and Uni...

In this paper, we explore spelling errors as a source of information for detecting the native language of a writer, a previously under-explored area. We note that character n-grams from misspelled words are very indicative of the native language of the author. In combination with other lexical features, spelling error features lead to 1.2% improvement in accuracy on classifying texts in the TOE...

With growing exposure of children to handheld and mobile devices, there is an increasing interest in exploring the use of mobile technology for educational purposes. In particular, touch-based devices seem to promise great potential in this domain. In this paper, we present Math Tutor – an Android-based application designed to help children learn and practice early numeracy addition and subtrac...

With the increasing number of mobile platforms available, developers aiming to bring their products to the larger number possible of consumers are required to program for different platforms, using different development environments, and often resulting in applications with different user experiences. A possible solution for this problem are development frameworks supporting develop once, deplo...

Attempts to profile authors based on their characteristics, including native language, have drawn attention in recent years, via several approaches using machine learning with simple features. In this paper we investigate the potential usefulness to this task of contrastive analysis from second language acquistion research, which postulates that the (syntactic) errors in a text are influenced b...

We outline the first application of Native Language Identification (NLI) to Finnish learner data. NLI is the task of predicting an author’s first language using writings in an acquired language. Using data from a new learner corpus of Finnish — a language typology quite different from others previously investigated, with its morphological richness potentially causing difficulties — we show that...