Suwen Intelligence, focusing on AI technologies for industry. We use AI and LLM based approaches to Collect, Filter, Label, and Deliver all key competitive intel for Strategy and Marketing departments, keep your company on top of the Chinese Market, technical trends, and high-tech talents, losing no valuable business deals.

AI scientist at Noah's Ark Lab, working on industry intelligence projects including: - NLP projects on question-answering and knowledge learning, applied on contract learning, bidding data learning, financial data learning- data clustering and prediction projects on Huawei's global supply chain

Working as a post-doc researcher with Stuart Russell at CNRS LIP6, collaborated with EECS department of UC Berkeley. Research on probabilistic models for NLP projects, such as knowledge representation with triples and knowledge inference with graphical models. Some particular efforts have been put to entity/mention models and also the convergence efficiency issues of Markov Chain Monte Carlo methods.

Phd research with Olivier Ferret and Romaric Besançon from CEA LIST and Brigitte Grau from CNRS-LIMSI of University Paris-Sud. Research on unsupervised open-domain information extraction in large scale. In particular, knowledge (or information) is represented as triples in the format "entity-relation-entity" (This kind of knowledge representation is named as Knowledge Graph by Google since 2012). My work includes how to discover valid relations (which is unknown in advance) between entities… Show more Phd research with Olivier Ferret and Romaric Besançon from CEA LIST and Brigitte Grau from CNRS-LIMSI of University Paris-Sud. Research on unsupervised open-domain information extraction in large scale. In particular, knowledge (or information) is represented as triples in the format "entity-relation-entity" (This kind of knowledge representation is named as Knowledge Graph by Google since 2012). My work includes how to discover valid relations (which is unknown in advance) between entities, and how to semantically group discovered relations according to their semantic similarities. Show less

Named Entity Recognition with statistic models, such as HMM, Maximum Entropy and CRF

Research and development of a class library for adaptive robot behaviors in Robotic Intelligence Lab