Expanding new functionalities efficiently is an ongoing challenge for single-turn task-oriented dialogue systems. In this work, we explore functionality-specific semi-supervised learning via self-training. We consider methods that augment training data automatically from unlabeled data sets in a functionality-targeted manner. In addition, we examine multiple techniques for efficient selection of augmented utterances to reduce training time and increase diversity. First, we consider paraphrase detection methods that attempt to find utterance variants of labeled training data with good coverage. Second, we explore sub-modular optimization based on n-grams features for utterance selection. Experiments show that functionality-specific self-training is very effective for improving system performance. In addition, methods optimizing diversity can reduce training data in many cases to 50% with little impact on performance.
Word embedding models offer continuous vector representations that can capture rich contextual semantics based on their word co-occurrence patterns. While these word vectors can provide very effective features used in many NLP tasks such as clustering similar words and inferring learning relationships, many challenges and open research questions remain. In this paper, we propose a solution that aligns variations of the same model (or different models) in a joint low-dimensional latent space leveraging carefully generated synthetic data points. This generative process is inspired by the observation that a variety of linguistic relationships is captured by simple linear operations in embedded space. We demonstrate that our approach can lead to substantial improvements in recovering embeddings of local neighborhoods.