Quantum Inspired Zeroshot Classification For Adverse Drug Reactions Detection From Social Media Reviews

A deep learning model enhanced with quantum transformer has been introduced in detection of adverse reactions of drugs within the reviews collected from social media written in English language. This proposed model enhanced the human language understanding for detecting adverse reactions of drug. Quantum transformers are used for position embedding and attention purpose. Variational quantum circuit encoded the human language and identify the relevant components of the input reviews as class feature vectors to develop a relationship network for each drug correspond to the condition and review comments on which probability distribution functions applied to detect ADRs found in a medicinal product or not. The correctness of advanced model architecture built on quantum transformer is quite encouraging, and it performs better than other cutting-edge learning models. Achieved an accuracy of 93% with a loss of 0.0393 with this proposed model. Though it is too early to declare that quantum computing has completely replaced. This proposed quantum enhanced transformer helps to show the direction towards quantum enhanced hybrid model design in research.