Learning to Classify Events from Human Needs Category Descriptions

Findings of the Association for Computational Linguistics: EMNLP 2020, pages 4698–4704 November 16 - 20, 2020. c⃝2020 Association for Computational Linguistics 4698 Learning to Classify Events from Human Needs Category Descriptions Haibo Ding and Zhe Feng Bosch Research and Technology Center, Sunnyvale, CA 94085, USA {haibo.ding,zhe.feng2}@us.bosch.com Abstract We study the problem of learning an event clas- sifier from human needs category descriptions, which is challenging due to: (1) the use of highly abstract concepts in natural language descriptions, (2) the difficulty of choosing key concepts. To tackle these two challenges, we propose LEAPI, a zero-shot learning method that first automatically generate weak labels by instantiating high-level concepts with pro- totypical instances and then trains a human needs classifier with the weakly labeled data. To filter noisy concepts, we design a reinforced selection algorithm to choose high-quality con- cepts for instantiation. Experimental results on the human needs categorization task show that our method outperforms baseline methods, producing substantially better precision. 1 Introduction Training accurate text classifiers often requires a large amount of manually labeled data, which is expensive to collect. In contrast, humans can often perform well on a classification task by only read- ing category descriptions, which are easy to obtain. It is desirable if computers can automatically learn classifiers from class descriptions. In this work, we aim to learn an event classifier automatically from unlabeled events by using human needs category descriptions as supervision. Human needs cate- gories have been proposed to explain why an event is positive or negative (Ding and Riloff, 2018a; Li and Hovy, 2017). For example, event “I had can- cer” is negative because it violates Health needs, while “I had steak” is usually positive because it matches Physiological needs. Human needs cate- gorization of events (Ding and Riloff, 2018a) is a task to classify events into eight categories asso- ciated with human needs (Maslow et al., 1970) in psychology: Physiological, Health, Leisure, Social, Finance, Cognition, Emotion, and None. Physiological+ Needs Description:+the+need+for+a+person+to+ obtain+food,+to+have+meals+… foodàfruit,+vegetable,+meat,+egg,+fish,… “I+bought+fruits”,+“I+had+eggs+this+morning” ConceptàInstances Labeled+Events Leisure+Needs Description:+the+need+for+a+person+to+ have+leisure+activities,+to+enjoy+art+… leisure/activitiesàfishing,+shopping,+golf “I+went+to+fishing”,+“Dad+went+to+play+golf” ConceptàInstances Labeled+Events Figure 1: Examples of human need descriptions, se- lected key concepts, and labeled events with prototypi- cal instances of key concepts. However, learning a classifier from human needs category descriptions directly is challenging. First, human needs category descriptions often consist of highly abstract concepts. As shown in Fig. 1, the Physiological and Leisure needs are defined using abstract concepts (e.g., “food”, “leisure activities”) to cover all instances of them. As demonstrated in our experiments, it is not easy to represent the meanings of these abstract concepts accurately us- ing existing methods. Second, it is not clear how to automatically choose key concepts without ac- cessing manual labels. In this work, we tackle these two challenges, and propose LEAPI, a method to automatically Learn a classifier from human need descriptions with Prototypical Instantiation. As shown in Fig.1, we first generate candidate key concepts from hu- man needs descriptions (e.g., “food”). Then we automatically assign human needs category labels to events that contain …

Content truncated to 500 words