Automatic Coding
Automatic coding, the automated generation and manipulation of code, aims to improve software development efficiency and reduce human error. Current research focuses on leveraging large language models (LLMs), often employing transformer architectures and agent-based frameworks, to translate natural language instructions into executable code, perform code analysis (including security assessments), and even automate program repair. This field is significant because it promises to increase developer productivity, enhance code quality and security, and potentially transform various domains, from medical record processing to robotic programming, through automation of coding tasks.
Papers
Extralonger: Toward a Unified Perspective of Spatial-Temporal Factors for Extra-Long-Term Traffic Forecasting
Zhiwei Zhang, Shaojun E, Fandong Meng, Jie Zhou, Wenjuan Han
Automatic programming via large language models with population self-evolution for dynamic job shop scheduling problem
Jin Huang, Xinyu Li, Liang Gao, Qihao Liu, Yue Teng