• Publications
  • Influence
Data from eye-tracking corpora as evidence for theories of syntactic processing complexity
It is concluded that eye-tracking corpora, which provide reading time data for naturally occurring, contextualized sentences, can complement experimental evidence as a basis for theories of processing complexity. Expand
Improving Variational Encoder-Decoders in Dialogue Generation
A separate VED model is developed that learns to autoencode discrete texts into continuous embeddings and generalize latent representations by reconstructing the encoded embedding through transforming Gaussian noise through multi-layer perceptrons. Expand
Information Presentation in Spoken Dialogue Systems
This work identifies compelling options based on a model of user preferences, and presents tradeoffs between alternative options explicitly, and suggests that presenting users with a brief summary of the irrelevant options increases users’ confidence in having heard about all relevant options. Expand
The time-course of processing discourse connectives
While concessive discourse markers can be processed rapidly if the context is constraining enough, there is a delay compared to causal contexts, and one way to investigate this issue is to focus on the time-course of discourse connectors. Expand
The Frequency of Rapid Pupil Dilations as a Measure of Linguistic Processing Difficulty
It is argued that the ICA is indicative of activity in the locus caeruleus area of the brain stem, which has recently also been linked to P600 effects observed in psycholinguistic EEG experiments. Expand
Modeling Semantic Expectation: Using Script Knowledge for Referent Prediction
It is found that script knowledge significantly improves model estimates of human predictions and the highly controversial hypothesis that predictability influences referring expression type is tested but does not find evidence for such an effect. Expand
Learning to predict script events from domain-specific text
The automatic induction of scripts (Schank and Abelson, 1977) has been the focus of many recent works. In this paper, we employ a variety of these methods to learn Schank and Abelson’s canonicalExpand
Event participant modelling with neural networks
Evaluation shows a drastic improvement over current state-of-the-art systems on modelling human thematic fit judgements, and it is demonstrated via a sentence similarity task that the system learns highly useful embeddings. Expand
Phonological Constraints and Morphological Preprocessing for Grapheme-to-Phoneme Conversion
It is shown that adding simple syllabification and stress assignment constraints, namely ‘one nucleus per syllable’ and ‘ one main stress per word’, to a joint n-gram model for g2p conversion leads to a dramatic improvement in conversion accuracy. Expand
A Systematic Study of Neural Discourse Models for Implicit Discourse Relation
To their surprise, the best-configured feedforward architecture outperforms LSTM-based model in most cases despite thorough tuning, and it is found that feedforward can actually be more effective. Expand