Potential project topics and supervisors
- This semester you have to provide the topic and supervisor by yourself.
- We do the grand introduction of topics and supervisors in autumn terms.
- Spring term projects are for those who really know what they want to do.
Semiautomatic terminology extraction tool
Student: Karl-Oskar Masing
Supervisor: Raul Sirel
Privacy-preserving Frequent Itemset Mining
Student: Alisa Pankova
Supervisor: Peeter Laud (Cybernetica)
Read recent articles about both privacy-preserving and ordinary frequent itemset mining. Based on these, think on a privacy-preserving protocol for finding frequent itemsets that seems the most suitable for data tables with sparse rows, where we do not attempt to hide the numer of rows/columns and number of non-zero entries in a row, but everything else should remain private. If there is time left, implement the algorithm on Sharemind and do benchmarking.
- A 10-20 page report, based on which the final presentation is made.
- An implementation on SecreC (if there is time) - it is probably not going to be a formal deliverable, and is interesting mostly because of the efficiency results.