معرفی
J. M. Howe is a Professor in the Department of Computer Science at City, University of London, with an active research career spanning over 25 years from 1997 to the present. With an ORCID identifier 0000-0001-8013-6941, Howe has established himself as a prominent researcher in formal methods, logic programming, and the intersection of symbolic and neural approaches to artificial intelligence.
Howe's research interests center on abstract interpretation, program analysis, constraint solving, and more recently, neural-symbolic AI. His early work focused on logic programming and abstract domains, particularly systems of two variables per inequality (TVPI), which has remained a consistent thread throughout his career. In recent years, he has expanded into explainable AI, rule extraction from neural networks, and the development of the Neural Multi-Space (NeMuS) framework for integrating symbolic and neural approaches. His research has practical applications in program verification, security (particularly cross-site scripting detection), and agricultural technology.
Analysis of Howe's publication trajectory reveals a consistent focus on foundational programming language theory that gradually evolved toward machine learning applications while maintaining strong theoretical underpinnings. His work shows increasing collaboration with researchers in neural networks while preserving his expertise in formal methods, creating a distinctive research niche at the intersection of symbolic and connectionist AI approaches.
Howe has made significant contributions to the understanding of widening operators in abstract interpretation, constraint solving techniques, and methods for extracting interpretable rules from black-box neural networks. His research demonstrates both theoretical depth and practical applicability across multiple domains including software verification, security analysis, and computer vision applications.
His collaborative work spans multiple institutions, with frequent collaborations with researchers like A. King, M. Brain, F. Mereani, and E. Robbins. This pattern of collaboration demonstrates his integration within both the formal methods and machine learning research communities.
