Made of Steel? Learning Plausible Materials for Components in the Vehicle Repair Domain
This addresses a domain-specific problem in vehicle repair for improving material prediction, but it is incremental as it adapts existing PLM methods.
The paper tackles the problem of learning plausible materials for vehicle repair components without annotated datasets by probing pretrained language models with cloze tasks and domain adaptation, achieving performance where a distilled PLM outperforms pattern-based algorithms and addressing data sparsity for 98% multiword components.
We propose a novel approach to learn domain-specific plausible materials for components in the vehicle repair domain by probing Pretrained Language Models (PLMs) in a cloze task style setting to overcome the lack of annotated datasets. We devise a new method to aggregate salient predictions from a set of cloze query templates and show that domain-adaptation using either a small, high-quality or a customized Wikipedia corpus boosts performance. When exploring resource-lean alternatives, we find a distilled PLM clearly outperforming a classic pattern-based algorithm. Further, given that 98% of our domain-specific components are multiword expressions, we successfully exploit the compositionality assumption as a way to address data sparsity.