Fuzzy Classification of Multi-intent Utterances
This work addresses the inherent vagueness in language and intent boundaries for natural language processing applications, particularly in task-oriented dialog systems, by introducing a novel fuzzy approach to classification.
The paper tackles the problem of ambiguity in intent classification for natural language utterances by proposing a fuzzy classification scheme that creates degree memberships over fuzzified intent classes, addressing both single-intent and multi-intent cases. It evaluates the approach on two task-oriented dialog datasets, revealing the impact of lexical overlap and data distributions on fuzzification, and compares defuzzified memberships to binary counterparts across different techniques.
Current intent classification approaches assign binary intent class memberships to natural language utterances while disregarding the inherent vagueness in language and the corresponding vagueness in intent class boundaries. In this work, we propose a scheme to address the ambiguity in single-intent as well as multi-intent natural language utterances by creating degree memberships over fuzzified intent classes. To our knowledge, this is the first work to address and quantify the impact of the fuzzy nature of natural language utterances over intent category memberships. Additionally, our approach overcomes the sparsity of multi-intent utterance data to train classification models by using a small database of single intent utterances to generate class memberships over multi-intent utterances. We evaluate our approach over two task-oriented dialog datasets, across different fuzzy membership generation techniques and approximate string similarity measures. Our results reveal the impact of lexical overlap between utterances of different intents, and the underlying data distributions, on the fuzzification of intent memberships. Moreover, we evaluate the accuracy of our approach by comparing the defuzzified memberships to their binary counterparts, across different combinations of membership functions and string similarity measures.