Having a good quality sleep is important for a person's well-being. This web application assesses people's sleep quality given EEG signal. First, it uses a sequence to sequence attention learning model for sleep stage classification. Then it calculates sleep stage transition rate and final score given the percentage of deep sleep.
Deep Web (e.g., Yelp, IMDb) is an invaluable external data source for enriching a local database with new attributes. ActiveDeeper is a model-driven data enrichment system powered by deep web. ActiveDeeper treats deep web as a labeler and uses it to train a multi-class classification model. It comes with a Google Sheets add-on supporting Yelp and Dblp demonstrating how results are generated and how users can check the support and confidence of results. The related paper is accepted by VLDB 2020 demonstration.
Software plagiarism has become a serious threat to the healthy working development of the software industry. Efﬁcient software plagiarism detection methods are of critical need.This detector focuses on university assignment plagiarism detection. It detects similarity of C programs by leveraging the characteristic of their basic blocks at LLVM Intermediate Representation(IR) level. It conducts a sampling-based dynamic analysis and incorporates the similarity of longest common subsequence of basic blocks.
This is a simple fault tolerance and isolation project based on LLVM. The project involves instrumenting an LLVM module in order to produce a new program with varying policies and strategies for fault tolerance. It protects programs from different types of common bugs by checking a few simple policies at runtime. It also has a few different policies for handling these types of faults. The different policies will allow users to control the quality of service (QOS) of the instrumented program in order to make a trade off between preservation of the original program semantics and availability. Checkout my code below for detailed policies.