BO LI
Senior Undegraduate
Audio Search
This project demonstrates a real-time search system called Aucher for live audio streams. It searches live audio streams by using keywords and voice, illustrates the trade-off of freshness, popularity and relevance on query results, performs searching hot terms in a time frame, and shows the ability of searching live audio streams in real-time.
View MoreAudio Search Demo
accepted by ICDE conference
- Date:
- Jan 2018 - May 2018
- Place:
- National University of Singapore
Audio streaming service has become increasingly popular with the wide use of smart phones. Because of the popularity of audio broadcasting, the data volume of live audio streams is also ever increasing. Searching and indexing these audio streams is an important and challenging problem. Aucher is a system prototype which can support both voice search and keywords search on audio streams. We achieved the real-time response for queries by novel index which exploits log structured merge-trees and supports multi-modal search. Moreover, it can handle insertion about four times faster and more memory efficient than the state-of-the-art solution.
Mid-infrared, Robust Retro-reflecting Tape
Inter-disciplinary Project: Computer Graphics & Material Science. This project built the first nano-scale retroreflector prototypes for collecting the environmental information of trace gases in the air.
View More
Mid-infrared, Robust Retro-reflecting Tape for Airbone Atmospheric Sensing
NASA Project
- Date:
- Jun 2018 - Aug 2018
- Place:
- Princeton University
People often use special aeroplanes to collect environmental information. This project built the first nano-scale retroreflector prototype used for detecting the amount and components of the trace gas in the air. The reflectors receive the light with mid-infrared energy wavelength and reflect it back to the open-path spectorscopic sensor. People can get the information of trace gases by comparing the difference between emitted light and reflected light. We used OpenGL to imitate the reflecting process and compute the reflective efficiency. Then we designed two retroreflector protypes using gray-scale photo-lithography: the full reflector and the truncated refletor, with their reflective effiency being 99% and 68% respectively.
Teaching Assistant AI Chatbot
This project aims at using intelligent chatbot technologies in innovative ways to improve the learning effectiveness of online education.
View MoreAI Chatbots for e-Learning Environments
- Date:
- June 2018 - May 2019
- Place:
- Hong Kong University of Science and Technology
Our AI chatbot acts as a teaching assistant for a an Open-edx course. The chatbot uses Rasa Stack to manage the dialog flow and applies some Natural Language Processing machine learning algorithms. It integrates bi-directoin GRU and attention model to detect similar questions and do the multi-intent classification. We built the index for Stack Overflow questions and answers. It can form a dialog with students and track the follow up questions.