Some actions must be executed in different ways depending on the context. For example, wiping away marker requires vigorous force while wiping away almonds requires more gentle force. In this paper we provide a model where an agent learns which manner of action execution to use in which context, drawing on evidence from trial and error and verbal corrections when it makes a mistake (e.g., ``no, gently''). The learner starts out with a domain model that lacks the concepts denoted by the words in the teacher's feedback; both the words describing the context (e.g., marker) and the adverbs like ``gently''. We show that through the the semantics of coherence, our agent can perform the symbol grounding that's necessary for exploiting the teacher's feedback so as to solve its domain-level planning problem: to perform its actions in the current context in the right way.

Domain classification is the task of mapping spoken language utterances to one of the natural language understanding domains in intelligent personal digital assistants (IPDAs). This is a major component in mainstream IPDAs in industry. Apart from official domains, thousands of third-party domains are also created by external developers to enhance the capability of IPDAs. As more domains are developed rapidly, the question of how to continuously accommodate the new domains still remains challenging. Moreover, existing continual learning approaches do not address the problem of incorporating personalized information dynamically for better domain classification. In this paper, we propose CoNDA, a neural network based approach for domain classification that supports incremental learning of new classes. Empirical evaluation shows that CoNDA achieves high accuracy and outperforms baselines by a large margin on both incrementally added new domains and existing domains.

We investigate a method for applying CBR to a source of data where there are no negative exemplars. Our problem domain is one of recommending characteristics of multidisciplinary collaborators based on a collection of funded grants. Thus, there are no negative exemplars. Lacking sufficient domain knowledge, we seek to apply a feedback algorithm to learn weights even in the absence of negative exemplars. Our approach is based on the assumption that well aligned cases, cases where similar problems have similar solutions, are better suited for learning feature weights. Our approach clusters the problem and solution spaces separately to identify well aligned cases. We also identify poorly aligned cases that may hinder effective learning of weights, and exclude them. The clusters of well aligned cases provide a means to utilize feedback algorithms. We use two methods, case alignment and case cohesion, to show that our approach succeeds in identifying well aligned cases. We also compare our approach to a method based on single class learning, a machine learning approach for reasoning without negatives. Our results show that our approach is viable to learning weight in the absence of negative exemplars.