My research interests lie in AI and NLP, particularly building practically effective and trustworthy natural language interfaces (NLIs, e.g., semantic parsers, question answering systems) to assist humans with knowledge acquisition and problem solving. Taskwisely, I am passionate to build NLIs to structured knowledge. Along this topic, I ground my research vision into three directions: (1) Constructing NLIs to computer programs, as exemplified by code generation/search/summarization tasks [WWW'18; WWW'19]; (2) Human-machine collaborative NLIs, where I investigate how humans and intelligent machines can communicate and collaborate to improve task completion and enhance system trustworthiness [AAAI'19; EMNLP'19]; (3) Learning NLIs from weak user supervision, where I am interested in constructing NLIs that can continually and efficiently improve themselves by learning from users [EMNLP'20].