In the new global economy, machine learning has become a central issue for most research fields as it offers techniques to tackle an extremely challenging real-world problem. The study of machine learning is important for addressing fundamental questions related to science and engineering. There are three subdomains of machine learning while; Supervised learning where training will take place only if the data has been labeled and consists of desired inputs and outputs, Unsupervised learning where the training data does not need any labeling and the environment only produces inputs without specific goals and finally learning by reinforcement where the characteristic of the information available in the training data lies between supervised and unsupervised and this type of learning occurs from feedback received through interactions with external environments. Say no to plagiarism. Get a tailor-made essay on "Why Violent Video Games Shouldn't Be Banned"? Get an original essay. Both supervised and unsupervised learning are suitable for data analysis, while reinforcement learning is suitable for handling problems involving decision making. With the rapid emergence of machine learning trends, the need to improve traditional machine learning into modern machine learning is very important. Improvement in terms of software (algorithm) and hardware is the key to achieving advanced machine learning that can handle current machine learning problems. In several advanced learning methods have been mentioned in order to improve traditional machine learning, including; Representational learning, deep learning, distributed and parallel learning, transfer learning, active learning, and kernel-based learning. Parallel learning is basically based on the parallel computing environment. Parallel computing defined as a set of interconnection processes between processing elements and memory modules. In machine learning, parallel computing has improved traditional machine learning by implementing the use of multicore processors instead of a single processor[2]. Some researchers have discussed and applied parallel computing to address machine learning problems. Qiu et. al. they come with a review paper that discusses how big data was processed using machine learning. As data becomes large and complex, traditional machine learning has difficulty training it. Therefore, as stated above, six advanced learning methods have been introduced. Next, five topics related to machine learning in big data were discussed. One of the problems includes understanding large-scale data. To solve this problem, a distributed framework based on parallel computing is suggested. Alternating Direction Method of Multipliers (ADMM), the framework that can produce algorithms with dispersion and scaling capabilities is very suitable. ADMM can break down multiple problems and helps identify the solution by coordinating those solutions into smaller groups of problems. Subsequently, the use of parallel programming methods to solve large-scale output datasets was also mentioned. Subsequently, Memeti et. al. review on two techniques for planning parallel computing systems which are Machine Learning and Meta-heuristic techniques. Since parallel computing is usually involved in very complex and resource-intensive problems, it is necessary. 231–234, 2017.