Neural-Symbolic Solver for Math Word Problems with Auxiliary Tasks

Qin, Jinghui; Liang, Xiaodan; Hong, Yining; Tang, Jianheng; Li, Lin

doi:10.48550/arxiv.2107.01431

Cited by 4 publications

(6 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Goal-driven decompositional tree-based decoders are a significant component in the stateof-the-art. Introduced in Xie and Sun (2019), this class of decoder is considered by all but three discussed models, as shown in Figure 5 and ex-tends to non-graph-based models (Qin et al, 2021;. In GTS, goal vectors guide construction of expression subtrees (from token node embeddings) in a recursive manner, until a solution expression tree is generated.…”

Section: Math Word Problemsmentioning

confidence: 99%

“…In GTS, goal vectors guide construction of expression subtrees (from token node embeddings) in a recursive manner, until a solution expression tree is generated. Proposed models do expand on the GTS-based decoder through the inclusion of semantically-aligned universal expression trees (Qin et al, 2020(Qin et al, , 2021, though this adaptation is not as widely used. The state-of-the-art Zhang et al, 2022) approaches follow the GTS decoder closely.…”

Section: Math Word Problemsmentioning

confidence: 99%

“…Language models that transfer knowledge learned from auxiliary tasks rival models based on explicit graph representation of problem text. As an alternative to encoding explicit relations through graphs, other work (Kim et al, 2020;Qin et al, 2021; relies on pre-trained transformer-based models, and those which incorporate auxiliary tasks assumed relevant for solving MWPs to latently learn such relations. However, it seems the case that auxiliary tasks alone do not deliver competitive performance (Qin et al, 2020) without the extensive pretraining efforts with large corpora, as we see with BERT-based transformer models.…”

Section: Math Word Problemsmentioning

confidence: 99%

See 2 more Smart Citations

A Survey in Mathematical Language Processing

Jordan¹,

Freitas²

2022

Preprint

View full text Add to dashboard Cite

Informal mathematical text underpins realworld quantitative reasoning and communication. Developing sophisticated methods of retrieval and abstraction from this dual modality is crucial in the pursuit of the vision of automating discovery in quantitative science and mathematics. We track the development of informal mathematical language processing approaches across five strategic sub-areas in recent years, highlighting the prevailing successful methodological elements along with existing limitations.

show abstract

Section: Math Word Problemsmentioning

confidence: 99%

Section: Math Word Problemsmentioning

confidence: 99%

Section: Math Word Problemsmentioning

confidence: 99%

See 1 more Smart Citation

A Survey in Mathematical Language Processing

Jordan¹,

Freitas²

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Auxiliary tasks could also be part of such a setup. Auxiliary tasks like Common Sense Prediction Task, Number Quantity Prediction, and Number location prediction could be helpful (Qin et al, 2021). The Commonsense prediction task aims to predict the commonsense required to solve an MWP, like the number of legs of a horse, or a number of days in a week, etc.…”

Section: Design Choicesmentioning

confidence: 99%

“…External knowledge could be commonsense world knowl-edge, particular domain knowledge, or complex mathematical formulas. There is some progress to infuse commonsense world knowledge or predict it using auxiliary tasks (Qin et al, 2021), however remaining two dimensions are yet to be addressed. Such external knowledge is incorporated using rule-based, pattern-matching approaches in non-neural models; on the other hand, neural models leverage data augmentation and auxiliary task training.…”

Section: Infusing Explicit and Definitive Knowledgementioning

confidence: 99%

Towards Tractable Mathematical Reasoning: Challenges, Strategies, and Opportunities for Solving Math Word Problems

Faldu¹,

Sheth²,

Kikani³

et al. 2021

Preprint

View full text Add to dashboard Cite

Mathematical reasoning would be one of the next frontiers for artificial intelligence to make significant progress. The ongoing surge to solve math word problems (MWPs) and hence achieve better mathematical reasoning ability would continue to be a key line of research in the coming time. We inspect non-neural and neural methods to solve math word problems narrated in a natural language. We also highlight the ability of these methods to be generalizable, mathematically reasonable, interpretable, and explainable. Neural approaches dominate the current state of the art, and we survey them highlighting three strategies to MWP solving:(1) direct answer generation, (2) expression tree generation for inferring answers, and (3) template retrieval for answer computation. Moreover, we discuss technological approaches, review the evolution of intuitive design choices to solve MWPs and examine them for mathematical reasoning ability. We finally identify several gaps that warrant the need for external knowledge and knowledgeinfused learning, among several other opportunities in solving MWPs.

show abstract

MWPToolkit: An Open-Source Framework for Deep Learning-Based Math Word Problem Solvers

Yihuai¹,

Wang²,

Lan³

et al. 2022

AAAI

View full text Add to dashboard Cite

While Math Word Problem (MWP) solving has emerged as a popular field of study and made great progress in recent years, most existing methods are benchmarked solely on one or two datasets and implemented with different configurations. In this paper, we introduce the first open-source library for solving MWPs called MWPToolkit, which provides a unified, comprehensive, and extensible framework for the research purpose. Specifically, we deploy 17 deep learning-based MWP solvers and 6 MWP datasets in our toolkit. These MWP solvers are advanced models for MWP solving, covering the categories of Seq2seq, Seq2Tree, Graph2Tree, and Pre-trained Language Models. And these MWP datasets are popular datasets that are commonly used as benchmarks in existing work. Our toolkit is featured with highly modularized and reusable components, which can help researchers quickly get started and develop their own models. We have released the code and documentation of MWPToolkit in https://github.com/LYH-YF/MWPToolkit.

show abstract

Neural-Symbolic Solver for Math Word Problems with Auxiliary Tasks

Cited by 4 publications

References 39 publications

A Survey in Mathematical Language Processing

A Survey in Mathematical Language Processing

Towards Tractable Mathematical Reasoning: Challenges, Strategies, and Opportunities for Solving Math Word Problems

MWPToolkit: An Open-Source Framework for Deep Learning-Based Math Word Problem Solvers

Contact Info

Product

Resources

About