
WikiReading: A Novel Largescale Language Understanding Task over Wikipedia
We present WikiReading, a largescale natural language understanding tas...
read it

LSOIE: A LargeScale Dataset for Supervised Open Information Extraction
Open Information Extraction (OIE) systems seek to compress the factual p...
read it

CAEHN: Commonsense Word Analogy from EHowNet
Word analogy tasks have tended to be handcrafted, involving permutations...
read it

Translating a Math Word Problem to an Expression Tree
Sequencetosequence (SEQ2SEQ) models have been successfully applied to ...
read it

LRW1000: A NaturallyDistributed LargeScale Benchmark for Lip Reading in the Wild
Largescale datasets have successively proven their fundamental importan...
read it

MathQA: Towards Interpretable Math Word Problem Solving with OperationBased Formalisms
We introduce a largescale dataset of math word problems and an interpre...
read it

DataDriven Methods for Solving Algebra Word Problems
We explore contemporary, datadriven techniques for solving math word pr...
read it
Ape210K: A LargeScale and TemplateRich Dataset of Math Word Problems
Automatic math word problem solving has attracted growing attention in recent years. The evaluation datasets used by previous works have serious limitations in terms of scale and diversity. In this paper, we release a new largescale and templaterich math word problem dataset named Ape210K. It consists of 210K Chinese elementary schoollevel math problems, which is 9 times the size of the largest public dataset Math23K. Each problem contains both the gold answer and the equations needed to derive the answer. Ape210K is also of greater diversity with 56K templates, which is 25 times more than Math23K. Our analysis shows that solving Ape210K requires not only natural language understanding but also commonsense knowledge. We expect Ape210K to be a benchmark for math word problem solving systems. Experiments indicate that stateoftheart models on the Math23K dataset perform poorly on Ape210K. We propose a copyaugmented and featureenriched sequence to sequence (seq2seq) model, which outperforms existing models by 3.2 of the Ape210K dataset. The gap is still significant between human and our baseline model, calling for further research efforts. We make Ape210K dataset publicly available at https://github.com/yuantiku/ape210k
READ FULL TEXT
Comments
There are no comments yet.