Recent progress through advanced neural models pushed the performance of task-oriented
dialog systems to almost perfect accuracy on existing benchmark datasets for intent …

Typically, spoken language understanding (SLU) models are trained on annotated data
which are costly to gather. Aiming to reduce data needs for bootstrapping a SLU system for a …

The lack of publicly available evaluation data for low-resource languages limits progress in
Spoken Language Understanding (SLU). As key tasks like intent classification and slot filling …

Typically, spoken language understanding (SLU) models are trained on annotated data
which are costly to gather. Aiming to reduce data needs for bootstrapping a SLU system for a …

Building conversational speech recognition systems for new languages is constrained by
the availability of utterances capturing user-device interactions. Data collection is expensive …

The lack of labeled training data for new features is a common problem in rapidly changing
real-world dialog systems. As a solution, we propose a multilingual paraphrase generation …

With synthetic data generation, the required amount of human-generated training data can
be reduced significantly. In this work, we explore the usage of automatic paraphrasing …

Content on the web is predominantly written in English, making it inaccessible to those who
only speak other languages. Knowledge graphs can store multilingual information, facilitate …

This paper presents an exploratory study that aims to evaluate the usefulness of
backtranslation in Natural Language Generation (NLG) from semantic representations for …

This paper addresses the question as to what degree a BERT-based multilingual Spoken
Language Understanding (SLU) model can transfer knowledge across languages. Through …