Evolutionary Computation
Our studies in evolutionary computation can be divided into three parts: platform, algorithm, and application. More specifically, with the support of highly efficient computing techniques in our platform, e.g., GPU-based speedups, we aim to develop a series of adaptive, scalable, and impactful evolutionary computation methods, to address diverse complexities presented in real-world applications, such as high-dimensionality, multi-level coupling, and computationally expensive evaluations, to mention a few. Click the picture for more information.
Machine Learning
At MIND, we build upon a solid foundation of large-scale and multimodal data to understand statistical regularities behind real-world data. From there, cutting-edge techniques such as transfer learning, federated learning, graph learning, automatic machine learning, and trustworthy machine learning empower us to unlock the full potential of our developed algorithms in realms like industrial 4.0, healthcare, material science, and beyond. Click the picture for more information.
Neuromorphic Computing
Learning from the brain’s information processing mechanism and system architecture represents a promising way to design more capable, efficient, robust, and explainable machines with general intelligence. The goals of our research will be to understand the neural coding, neural plasticity, and organization principles underlying biological neural systems and, ultimately, to develop low-power, robust, adaptive, and explainable neuromorphic cognitive machines. Click the picture for more information.
AI for Healthcare
Our research harnesses advanced AI technologies to revolutionize healthcare. By leveraging medical images and high-throughput sequencing data, we empower healthcare professionals with unprecedented accuracy for diagnoses, personalized treatments, and real-time monitoring. Amid this remarkable progress, we confront challenges of data quality, diversity, privacy, and transparency. Driven by a passion to overcome these challenges, we aspire to unlock AI's potential, democratizing better healthcare for all. Click the picture for more information.
AI for Material Science
Our interdisciplinary team targets for the discovery and optimization of high-performance materials, aiming to help solve urgent energy challenges. To achieve these goals, we have adopted an AI-driven paradigm to accelerate the screening and generation of novel materials. To fully scrutinize the potential of our developed high-performance materials and achieve the ultimate practical deployment, we will actively engage with academic and industrial collaborators specializing in catalyst synthesis and characterization and relevant device fabrications. Click the picture for more information.
