Interview with Dr. Lv Zhengdong, Founder & CTO of DeepCuriosity on AI and Natural Language Processing

At the 2018 AI Pioneer Conference in Hangzhou, the DataFun community interviewed Dr. Lv Zhengdong, founder and CTO of DeepCuriosity.

Dr. Lv holds a Ph.D. in computer science from the United States, serves as chief expert at the Xinjiang Public Safety Laboratory, and is recognized internationally for his work in deep learning, especially natural language processing.

In early 2013 he established the deep‑learning team at Huawei’s Noah’s Ark Lab, building software and hardware platforms from scratch and leading the lab to become a world‑class research institute in neural language intelligence within two years.

In 2016 he founded DeepCuriosity, applying cutting‑edge technologies such as neural‑symbolic models to the legal, public security, and financial sectors, and introduced the Object‑Oriented Neural Planning (OONP) framework for complex document understanding, which received high praise from both academia and industry.

His team’s work was highlighted in a 2017 AI Journal review of ten major contributions from Greater China, with four of those contributions authored by Dr. Lv.

He is also an inventor on several deep‑learning‑based NLP patents covering semantic matching, question answering, multi‑turn dialogue, and automated SMS reply.

During the interview, Dr. Lv explained that his interest in natural language processing began when he realized deep learning could be used to design architectures and mechanisms for understanding language, a field he sees as crucial to AI because it is close to logic, everyday life, and core AI tasks like reasoning and knowledge representation.

He described AI as the core of DeepCuriosity’s business, with applications focused on natural language understanding for public security, finance, and legal domains, aiming to assist experts in decision‑making and dramatically improve efficiency.

DeepCuriosity’s central idea is the deep integration of symbolic intelligence and neural networks—"neural‑symbolic"—which has been successfully deployed in court and financial settings; however, challenges remain such as the technology not yet reaching a practical usefulness threshold and data silos that limit NLP deployment.

His advice to newcomers is to balance breadth with depth and theory with hands‑on practice.

He noted that real‑world projects often face dirty data and poorly defined tasks, so breaking problems into well‑defined sub‑tasks and identifying core technical barriers is essential.

Regarding AI’s development curve, Dr. Lv observes that unrealistic hype is fading, leading to the collapse of companies lacking core technology or data, while academic interest is surging and research is becoming more diverse.

He urges professionals to engage with the impending societal transformation driven by AI rather than be displaced by it.

At the conference, he will present DeepCuriosity’s latest breakthroughs in natural language understanding and demonstrate how these “black‑tech” innovations are applied across various industry‑specific language‑intelligent products.

Note: Dr. Lv will share a talk titled “From Language Intelligence to Industry Intelligence” at the NLP track of the AI Pioneer Conference.

Images of the event and a friendly request to share the content conclude the article.