Generating Math Word Problems from Equations with Topic Consistency Maintaining and Commonsense Enforcement

Abstract

Recent years have seen significant advancement in text generation tasks with the help of neural language models. However, there exists a challenging task: generating math problem text based on mathematical equations, which has made little progress so far. In this paper, we present a novel equation-toproblem text generation model. In our model, 1) we propose a flexible scheme to effectively encode math equations, we then enhance the equation encoder by a Varitional Autoencoder (VAE) 2) given a math equation, we perform topic selection, followed by which a dynamic topic memory mechanism is introduced to restrict the topic distribution of the generator 3) to avoid commonsense violation in traditional generation model, we pretrain word embedding with background knowledge graph (KG), and we link decoded words to related words in KG, targeted at injecting background knowledge into our model. We evaluate our model through both automatic metrices and human evaluation, experiments demonstrate our model outperforms baseline and previous models in both accuracy and richness of generated problem text.

DISK: Domain-constrained Instance Sketch for Math Word Problem Generation

“…So this method is time-consuming and limited. Cao et al (2021) incorporates topic controlling and commonsense enforcement in MWP generation. Data-to-text Generation: Data-to-text generation transforms structured data into descriptive texts (Siddharthan, 2001;Gatt and Krahmer, 2018).…”

Math Word Problem Generation with Multilingual Language Models

“…A Variational Auto-Encoder (VAE) is used to generate the MWP from this encoding. Cao et al (2021) also make use of a VAE to bridge the gap between abstract math tokens and text. In addition to the equation and common sense knowledge graph as input, they take in the question text, as well as a set of words representing a topic.…”

“…Early solutions to MWP generation relied on template-based approaches (Polozov et al, 2015), and question rewriting (Koncel-Kedziorski et al, 2016). More recently, Recurrent Neural Networks (RNN) (Zhou and Huang, 2019;Liyanage and Ranathunga, 2020), fine-tuning pre-trained language models as well as Variational Autoencoders (VAE) Cao et al, 2021) have been proposed. Only Liyanage and Ranathunga (2020) have tried their NN solution in a multilingual setting, however the results are sub-optimal.…”