This is the code for the paper http://www.aclweb.org/anthology/I17-3017
Blog link https://medium.com/@purvanshimehta/can-computers-learn-maths-through-deep-learning-7ef714ca2cb3
Takes a question with two operands as input and predicts the operation between them. This system presents a novel approach to solve simple arithmetic word problems.We predict the operation that is to be performed (’-’,’+’,’*’,’/’) through a deep neu- ral network model and then use it to generate the answer.
The question is divided into two parts - worldstate and query as shown in Figure .
The worldstate and the query are processed separately in two different networks and finally the networks are merged to predict the final operation.
Our system learns to predict operationswith 88.81 % in a corpus of primary school questions.
pip install -r requirements.txt
- Testing on our data
python3 model.py DATA/train_LSTM_26112016 DATA/test_LSTM_26112016
run python3 model1.py
The function main_function returns the answer
API code is in main.py
- Testing on ARIS
python3 test_aris.py
You can change the input question eg -
-
Joan found 70 seashells on the beach . she gave Sam some of her seashells . She has 27 seashell . How many seashells did she give to Sam ?
-
There were 28 bales of hay in the barn . Tim stacked bales in the barn today . There are now 54 bales of hay in the barn . How many bales did he store in the barn ?
-
Mary is baking a cake . The recipe wants 8 cups of flour . She already put in 2 cups . How many cups does she need to add ?
-
Sara 's high school played 12 basketball games this year . The team won most of their games . They were defeated during 4 games . How many games did they win ?