Structured Language Generation Model for Robust Structure Prediction

arXiv:2402.08971v1 Announce Type: new
Abstract: We propose Structured Language Generation Model (SLGM), a mixture of new loss function and inference method for better generalization of structured outputs. Previous studies on structure prediction (e.g. NER, RE) make use of explicit dataset information, which would boost performance, yet it might pose challenges to robust generalization in real-world situations. Instead, our model gives generalized format information about data indirectly. With format information, we could reduce sequence-to-sequence problem into classification problem via loss calibration …

abstract arxiv boost challenges cs.cl dataset function inference information language language generation loss ner performance prediction robust studies type