Parallel Big Bang-Big Crunch-LSTM Approach for Developing a Marathi Speech Recognition System

Table 1

Work done in the proposed area.

Author [citation]	Techniques used	Characteristics	Issue faced
Smit et al. [8]	RNN	It increases the performance with a better accuracy rate	This model is not suitable for a large amount of learning data
Tu et al. [9]	DNN and unidirectional LSTM	Reducing word error rate (WER)	This model is not suitable for executing the objective functions with joint learning
Kipyatkova and Karpov [10]	Artificial neural network	It reduces WER	However, this model suffers from the demographic influence on the languages
Zhou et al. [11]	mDNN	It increases the recognition performance and increases the training speed	Conversely, the accuracy rate can be degraded
Xue et al. [12]	DNN	(i) It improves the performance and efficiency while adapting larger DNN models	This model cannot optimize the speaker representations
		(ii) It attains less WER