Master theses
Automatic personalized workflows
Keywords: Linked Data, Semantic Web, Personal Data, Reasoning, Workflows
Promotors: Pieter Colpaert, Ben De Meester
Students: max 1
Problem
Every organization needs to cater to an enormeous amount of processes and workflows. For example, within the imec FAB there are hundreds of chips-making recipes going on pretty much all at the same time, where one mistake could mean the loss of millions of dollars. More close to home, think of all the administrative hoops you need to jump to move houses.
Within KNoWS, we are working on a workflow composer (https://github.com/KNowledgeOnWebScale/oslo-steps-workflow-composer): an engine that, given a desired goal, generates a couple of alternative paths to reach that goal. Moreover, this engine can take limited context into account, for example, based on the steps you have previously taken, it won’t propose the same step twice.
Goal
In this thesis, try to improve the functionality of our workflow composer. There are multiple ways to tackle this:
- Improve context support, such as automatically extracting your relevant profile info (e.g., your birthdate) to further personalize and optimize the suggested paths.
- Add functionality, such as automatic calculations
- Improve performance (if a routeplanner can handle thousands of nodes, our workflow composer should handle thousands of steps)
- etc.