ARTIFICIAL INTELLIGENCE IN PHARMA PROCESSING

In 2002, certain amendments were proposed by food and drug administration in cGMP (current good manufacturing process) to modernize and improve the regulations for drug manufacturing and quality. Considering the amendments International Conference of Harmonization (ICH) developed Q8 guideline for pharmaceutical development which is based on the concept of Quality by Design (QbD). QbD is a systematic approach to pharmaceutical development with pre-defined objectives of designing manufacturing processes and developing manufacturing formulation of prescribed quality and offers better understanding of critical process. Later Q9 (Quality Risk Management) and Q10 (Pharmaceutical Quality System) guideline was introduced based on Q8. Pharmaceutical manufacturing is a complex process including multi variable interaction between process condition and raw material which is important for processing and defining the quality of the end product. This developed an urge in the researchers of employing design of experiment (DoE) to link CQAs (Critical Quality Attributes) to process parameters and API and excipient attributes and to define the design space. Artificial Intelligence can establish relationship between process parameters and various formulations in more understandable way while saving huge amount of money and time. Genetic algorithms, neural network are the technologies growing rapidly in pharmaceutical quality control processing. Artificial neural network is a data processing and learning machine which is inspired by the functioning of human brain. Human brain works slow but potential to perform complex tasks that computer is unable or might take long time to perform. ANN consists of artificial neurons (connected parallel) and weighs also called synaptic strength that helps to emulate the nervous system. Neural networks have capability of establishing the relation between the input and output data without prior assumptions or knowledge about the given data set that makes it suitable for regression and classification tasks which plays a vital role in many biomedical applications. Conventional methods for analysing the data is linear but neural network are non-linear (inherently) that makes it practicable for preparing models of complex data accurately. Hence, ANN can be used to solve many problems in biomedical and healthcare sector. This review article aims on demonstrating the successful and innovative application of Artificial Neural Network in the pharmaceutical industries.