Scientists develop automated machine learning system for biological research

2023-07-10

In a groundbreaking study, researchers at the Massachusetts Institute of Technology (MIT) have developed an automated machine learning system called BioAutoMATED that generates AI models for biological research. Led by Dr. Jim Collins and Professor Timothy Lu in the field of medical engineering and science, this innovative system aims to simplify the process of building machine learning models for scientists and engineers in the field of biology. Not only can this system select and construct appropriate models for a given dataset, but it can also handle tedious data preprocessing tasks. By reducing the time and effort required, BioAutoMATED offers new possibilities for biological science researchers. Recruiting machine learning experts can be a time-consuming and expensive process for scientific and engineering laboratories. Even with experts involved, the selection process still requires a significant amount of time and effort to choose the right models, format datasets, and fine-tune the models. According to Google's training course on machine learning fundamentals, data preparation and transformation alone can account for 80% of project time. This obstacle often discourages researchers from utilizing machine learning techniques in biology. BioAutoMATED is an automated machine learning system designed specifically for biological research. While automated machine learning (AutoML) systems are still relatively new, with most applications focused on image and text recognition, BioAutoMATED extends the capabilities of AutoML to the field of biological sequences. This is crucial because the fundamental language of biology is based on sequences such as DNA, RNA, proteins, and carbohydrates. One key advantage of BioAutoMATED is its ability to explore and construct various types of supervised machine learning models, including binary classification models, multi-class classification models, and regression models. By integrating multiple tools into one system, BioAutoMATED provides a larger search space than individual AutoML tools, making model selection more flexible and accurate. Traditionally, conducting experiments at the intersection of biology and machine learning has been an expensive task. Research teams often need to invest in a significant amount of digital infrastructure and trained manpower before determining the feasibility of their ideas. The goal of BioAutoMATED is to reduce these barriers, allowing researchers to freely conduct preliminary experiments and assess the feasibility of further experiments. This way, they can determine whether it is necessary to hire machine learning experts to build different models for their research. The benefits of using BioAutoMATED are manifold. Firstly, it greatly reduces the time and effort required to build AI models for biological research. What used to take weeks of work can now be completed in a matter of hours. This time-saving effect allows researchers to focus more on their core research objectives rather than getting caught up in the technical details of machine learning. Secondly, BioAutoMATED is particularly suitable for research teams with smaller biological datasets. It can explore models that are better suited for such datasets, as well as more complex neural networks. This versatility ensures that researchers can make full use of the data they have and gain meaningful insights. To promote application and collaboration, researchers have made the code of BioAutoMATED publicly available on GitHub. They encourage others to improve their work and collaborate with the larger community to develop BioAutoMATED into a tool that serves everyone. By raising awareness and integrating biological practices with the fast-paced AI-ML practices, the field of biological research can be advanced. BioAutoMATED represents a significant breakthrough in the field of biological research. Through the automated generation of AI models, this innovative system enables scientists and engineers to leverage machine learning in their research. By selecting appropriate models and handling data preprocessing, BioAutoMATED simplifies the research process and lowers the entry barriers for biological science researchers. With the continuous development of this field, the possibilities for collaboration and exploration are limitless. Source: https://readwrite.com/scientists-develop-automated-machine-learning-system-for-biology-research/

RECENT AI NEWS

RECENT AI TOOLS