Open World Classification with Adaptive Negative Samples
This addresses the problem of open world classification in NLP, which is crucial for real-world applications where unknown categories can emerge, and it represents an incremental advance by improving decision boundaries without prior knowledge.
The paper tackles the challenge of open world classification, where unknown categories appear only during inference, by proposing an adaptive negative samples (ANS) approach that generates synthetic open category samples during training without external data. The method achieves significant improvements over state-of-the-art methods on three benchmark datasets.
Open world classification is a task in natural language processing with key practical relevance and impact. Since the open or {\em unknown} category data only manifests in the inference phase, finding a model with a suitable decision boundary accommodating for the identification of known classes and discrimination of the open category is challenging. The performance of existing models is limited by the lack of effective open category data during the training stage or the lack of a good mechanism to learn appropriate decision boundaries. We propose an approach based on \underline{a}daptive \underline{n}egative \underline{s}amples (ANS) designed to generate effective synthetic open category samples in the training stage and without requiring any prior knowledge or external datasets. Empirically, we find a significant advantage in using auxiliary one-versus-rest binary classifiers, which effectively utilize the generated negative samples and avoid the complex threshold-seeking stage in previous works. Extensive experiments on three benchmark datasets show that ANS achieves significant improvements over state-of-the-art methods.