COMS 4732 Computer Vision II - Prof. Carl Vondrick - Spring 2022

As a research assistant at Westlake University, I do both developing a system as well as do some researches.For the system, I using the rasa frameworks to build a reservation chatbot for clients to make reservations with mental therapists or talk to our AI chatbot.For the other, I employed machine learning techniques, especially reinforcement learning, to enhance the quality of generated text by our AI chatbot.

Work as a research intern at K2data.

Key Achievements:

• Building a crawl to collect daily steel prices from the open-source website.With my system, more than 300 pieces of information on different websites become easily together.

• Use the LSTM model to make predictions for future steel prices, the accuracy of the testing data could achieve 83%.These prediction results made by my model could provide some help for the data analysis of field experts.