Data Analysis Software Project for Natural Language
TUCaN: 20-00-0948-pp / Moodle: DASP2020
The course will be taught in English. If you have any questions regarding this project, please contact Steffen Eger.
IMPORTANT NOTE: The course will take place in accordance with the adjusted schedule and modalities for the summer term (Covid-19 pandemic effects). Thus, the below dates are subject to change.
21.04.2020- Kickoff/ Presentation of project ideas (09:50-11:30; online) UKP researchers will present the project ideas, and the requirements for this course will be discussed.
26.04.2020 - Deadline for project selection Students select their preferred projects
28.04.2020 – Topic assignment (will be announced via moodle) We assign students to projects according to their preferences
30.06.–14.07.2020 – Project presentations by students (Tuesdays, 09:50-11:30)
In this software project, the students can choose from a wide range of different topics from several NLP areas, e.g., Argumentation Mining, Metaphor Detection, NLP for Social Good, and more. Our website provides additional information on which topics might be available (the actual project ideas will be presented in the first lecture).
The students' task in this course will be to execute one of the projects in small groups. This involves, for example, collecting and processing large datasets from the web, training deep neural networks and making them more efficient, deploying trained models, visualizing and analyzing neural network internals, and creating prototypes.
Students will work on the chosen projects in small groups. They will be mentored by one UKP researcher, who will define the project requirements. Students will regularly meet with their mentor to discuss the progress of the project, and they are required to hand-in short status reports to the teaching staff.
- Interest in working with natural language textual data
- Programming skills (most projects are going to use Python)
Steffen Eger (Please contact by e-mail for an appointment)
Prof. Dr. Iryna Gurevych