2016-2017 M.Sc Projects offered by Prof. C.L. Wang

These projects are offered for M.Sc(CS) students enrolled in Sept 2016.

Title (1): Cloud-based Peer-to-peer Marketplace for Small Services

Description:

An online peer-to-peer marketplace brings people and/or businesses together online to deal with each other directly without having to go through a slow, expensive middle man. The peer-to-peer marketplace not only saves time for the consumers by having everything in one place, but also allows them to choose and select services with ease 24 hours a day. In recent years, we have seen tremendous innovation in the field of building and running marketplaces, including accommodation (AirBnB),  transportation (Uber), on-demand household services (TaskRabbit), collaborative finance (Lending Club).  The project involves the development of two important components: (1) a demand/supply-driven matchmaking service to provide a match between a peer owning a certain resource and a peer in need of that resource, at the right time (when both the offering and need still exist) and against reasonable transaction costs (e.g., physical distance between both consumers cannot be too high).  (2) a user-based reputation system for quality control, ensuring a level of trust between consumers and service providers who have not previously met.  We will use Apache Spark for collecting and analyzing recommendations, reviews, and ratings.


Number of students if team project (max 3 students): 1-3
Stream: General
Prerequisites if any: The students who want to take this project should take COMP7305 offered in the 2nd semester.

Title (2):  SC-Redis: In-memory NoSQL with Strong Consistency Support

Description:

Many NoSQL storage systems compromise consistency in favor of availability and speed, by offering a concept of "eventual consistency" in which database changes are propagated to all nodes "eventually" (typically within milliseconds). Under eventual consistency model, however, queries for data might not return the most updated data immediately or might result in reading data that is not accurate. This project aims at enhancing Redis, an in-memory NoSQL database, with strong consistency and transactional correctness guarantees. Such strong replica consistency implies that once an object value has been changed by a write operation, all subsequent reads to that object will immediately see the the new value. The project work involves the modification of Redis (written in C) and the implementation of the Raft consensus protocol (https://raft.github.io/).

Number of students if team project (max 3 students): 1-3
Stream: General
Prerequisites if any: This is a research-oriented project. The students who want to take this project should have some system programming experiences or take COMP7305 offered in the 2nd semester.

Title (3): ThumbDL: Deep Learning Using Embedded/Integrated GPUs

Description:

This project is to optimize the performance of various Deep Learning models (e.g., ConvNets (AlexNet), RNNs (LSTM), Autoencoders) to offer a competitive energy-efficient solution for lightweight deep learning computations on embedded and integrated GPUs, such as Intel Iris Pro Graphics and Qualcomm Adreno GPUs. The ultimate goal is to make the traditional Deep Learning tool (e.g., Caffee) be able to run on mobile client efficiently for surveillance cameras and drones, robots, and AR/VR headsets for gesture and eye tracking. The OpenCL programming model will be used for the project. 

Number of students if team project (max 3 students): 1-2
Stream: General
Prerequisites if any: This is a research-oriented project. The students who want to take this project require some background of the GPU hardware architecture.

2016-2017 M.Sc Projects

2015-2016 M.Sc  Projects