Table-to-Text Natural Language Generation with Unseen Schemas

TL;DR

This paper introduces a new task and benchmark for table-to-text natural language generation focusing on unseen schemas, proposing a model that aligns unseen schemas to seen ones to improve generalization.

Contribution

It defines the task of NLG with unseen schemas, creates a benchmark dataset, and proposes a schema alignment model to enhance generalization to unseen attribute types.

Findings

Our model outperforms baselines significantly on the new benchmark.

Unseen schema alignment improves text generation quality.

The task presents unique challenges compared to standard data-to-text settings.

Abstract

Traditional table-to-text natural language generation (NLG) tasks focus on generating text from schemas that are already seen in the training set. This limitation curbs their generalizabilities towards real-world scenarios, where the schemas of input tables are potentially infinite. In this paper, we propose the new task of table-to-text NLG with unseen schemas, which specifically aims to test the generalization of NLG for input tables with attribute types that never appear during training. To do this, we construct a new benchmark dataset for this task. To deal with the problem of unseen attribute types, we propose a new model that first aligns unseen table schemas to seen ones, and then generates text with updated table representations. Experimental evaluation on the new benchmark demonstrates that our model outperforms baseline methods by a large margin. In addition, comparison with standard data-to-text settings shows the challenges and uniqueness of our proposed task.