Advanced Topics in Knowledge Management and Discovery KMD

News:

Presence: "Onlineveranstaltung – Mix aus synchron und asynchron"

Timetable

Day	Time	Frequency	Room	Lecturer	Max. participants
Seminar - Seminar - Dates/Times/Location:
Mon.	13:00 bis 15:00	weekly	G29-021	Spiliopoulou	20
Mon.	13:00 bis 15:00	weekly		Unnikrishnan	20

Overview (from LSF)

Learning Content	UPDATE 06.04.2021 The course materials are disseminated through Moodle: Click here to view the moodle course In light of the COVID-19 pandemic and shifting research focus towards leveraging data science methods to better understand the dynamics of the pandemic, the KMD seminar will be offered in a different format this semester. The new format will provide some materials as a starting point and give students the opportunity to use this (if desired by the student, combined with other data sources) to come up with their own research questions, find existing literature, and work together to answer questions that the students' propose to answer. The solutions will then be implemented by the students and then evaluated to test the efficacy of their method. i.e., The seminar will involve the entire data science pipeline starting at a data source that the students may choose to augment, coming up with interesting research questions that will be peer-reviewed by other seminar participants, and then also implemented and tested by the students in separate groups. The early ideation process where students will propose problems that they will solve will happen in a form analogous to an "exercise session", where students propose novel solutions, and iteratively improve the problem specification as well as data collection and evaluation techniques. Once the teams have an early-approved version of their research question, this will be presented to Prof. Spiliopoulou for additional feedback. Once the research questions are fixed, the students will hydrate the public domain twitter dataset to collect data that is releavant to their research question (and also collect any further data if applicable), implement and evaluate the solutions that they propose. Students are advised that the format of the seminar is changed considerably from the previous semesters, and comes with an implicit requirement that the seminar also needs that students are reasonably comfortable with implementing the algorithms that are necessary to solve the questions they propose. Python is preferred. At the end of the seminar, the results should be submitted in form like jupyter notebooks, so that they can be viewed, shared and disseminated more easily.
Literature	Recent publications on advanced topics in knowledge management and discovery (data mining, machine learning, ...)

Learning Content

UPDATE 06.04.2021 The course materials are disseminated through Moodle: Click here to view the moodle course

In light of the COVID-19 pandemic and shifting research focus towards leveraging data science methods to better understand the dynamics of the pandemic, the KMD seminar will be offered in a different format this semester. The new format will provide some materials as a starting point and give students the opportunity to use this (if desired by the student, combined with other data sources) to come up with their own research questions, find existing literature, and work together to answer questions that the students' propose to answer. The solutions will then be implemented by the students and then evaluated to test the efficacy of their method. i.e., The seminar will involve the entire data science pipeline starting at a data source that the students may choose to augment, coming up with interesting research questions that will be peer-reviewed by other seminar participants, and then also implemented and tested by the students in separate groups.

The early ideation process where students will propose problems that they will solve will happen in a form analogous to an "exercise session", where students propose novel solutions, and iteratively improve the problem specification as well as data collection and evaluation techniques. Once the teams have an early-approved version of their research question, this will be presented to Prof. Spiliopoulou for additional feedback. Once the research questions are fixed, the students will hydrate the public domain twitter dataset to collect data that is releavant to their research question (and also collect any further data if applicable), implement and evaluate the solutions that they propose.

Students are advised that the format of the seminar is changed considerably from the previous semesters, and comes with an implicit requirement that the seminar also needs that students are reasonably comfortable with implementing the algorithms that are necessary to solve the questions they propose. Python is preferred.

At the end of the seminar, the results should be submitted in form like jupyter notebooks, so that they can be viewed, shared and disseminated more easily.

Literature

Recent publications on advanced topics in knowledge management and discovery (data mining, machine learning, ...)