Solving Arithmetic Word Problems Using Block Recurrent Transformer

Recent improvements in deep learning techniques have shown promising results in solving arithmetic word problems. The Block Recurrent Transformer (BRT) is a neural network architecture that coagulates the benefits of transformers and recurrent neural networks (RNNs). This paper investigates the use of BRT for arithmetic problem-solving and compares its performance with other state-of-the-art methods. Furthermore, a detailed description of the methodology that evaluates the performance of BRT on a dataset of arithmetic word problems is given. Our results show that BRT outperforms other methods and reaches state-ofthe-art performance.