Computer Science Graduate Seminar: Operation-Based Model Recommenders

Monday, November 06, 2017, 2:30pm

Location: E2, Room 5055, Ahornstraße 55

Speaker: Dipl. Inform. Andreas Ganser

Abstract:

State of the art integrated development environments provide support for programmers in numerous ways taking cognitive load of programmer's backs. Often the basis for that lies in reusing knowledge or experience to gain task-tailored recommendations. For modelers this is missing for there is no common basis on how to identify and extract beneficial parts, how to warehouse and organize them, how to change and improve them, as well as how to query for and retrieve them for reutilization. Here we approach these challenge with operation-based recommenders and their realization called HERMES in a fully integrated manner.

This brings about a foundation for transferring approaches from recommendation systems as commonly known to modeling tasks. As a result, reuse of modeled knowledge can be widely automated for modelling to an extent providing support without patronizing any modeler. Further, we derive schema for recommendation production from our deployment experiences and as we take into account monitored contextual information, we could conclude that we got an operation-and-knowledge-based recommender system which supports (semi-)automating tedious activities as reasonable.

The computer science lecturers invite interested people to join.