Bilder vom Neuenheimer Feld, Heidelberg und der Universität Heidelberg

Lehrveranstaltungen
heiCO
Ressourcen	Fachschaft
Studien-FAQ	Technik-FAQ

NLP in Industry: challenges and best practices

Module Description

Course	Module Abbreviation	Credit Points
BA-2010[100%\|75%]	CS-CL	6 LP
BA-2010[50%]	BS-CL	6 LP
BA-2010	AS-CL	8 LP
Master	SS-CL, SS-TAC	8 LP

Lecturer	Daniel Dahlmeier
Module Type	Proseminar / Hauptseminar
Language	English
First Session	29.10.2021
Time and Place	Friday, 09:15-10:45, Online
Commitment Period	tbd.

Prerequisite for Participation

Mathematical Foundations of CL (or a comparable introductory class to linear algebra and theory of probability)
Programming I (Python)
Statistical Methods for CL (or a comparable introductory class to machine learning)

Assessment

Regular and active attendance of seminar (40%)
Project: novelty of idea, empirical results, clarity of report and presentation (60%)

Content

This seminar focuses on common challenges and best practices for natural language processing (NLP) in industry. NLP in industry comes with its own challenges, like data sparsity, data privacy regulations, and cost-benefit trade-offs. In this seminar, we discuss these challenges and technical approaches to overcome them. We will read and discuss recent industry-related publications, in particular from the NAACL 2021 industry track. Every student will get to present one paper in the seminar and write a short report on the presented topic.

Module Overview

Agenda

Date

Session

Materials

Literature

Daniel Jurafsky, James H. Martin. 2008. Speech and Language Processing (2nd Edition)
Alfred Spector, Peter Norvig, and Slav Petrov. 2012.Google's hybrid approach to research. Communications of the ACM 55(7).
Daniel Dahlmeier. 2017. On the Challenges of Translating NLP Research into Commercial Products. In Proceedings of ACL
D. Sculley, Gary Holt, Daniel Golovin, Eugene Davydov, Todd Phillips, Dietmar Ebner, Vinay Chaudhary, and Michael Young. 2014. Machine learning: The high interest credit card of technical debt. In Proceedings of SE4ML: Software Engineering for Machine Learning (NIPS 2014 Workshop)