729A97 Language Engineering Systems / 2016 Introduction Marco Kuhlmann Department of Computer and Information Science What will you do in this course? • survey a research field within natural language processing topic: universal dependencies • plan, implement and evaluate a project not necessarily on universal dependencies! • write scientific papers on universal dependencies, on your project work • get acquainted with standard practices in the field reviewing, where to find literature, form of a project proposal Universal Dependencies • framework for grammatical annotation • transparent and consistent across languages • work in progress for 28 languages • wide range of interesting applications multilingual parsing research and development, cross-lingual learning and annotation projection, empirical language typology Structure of the course • Reading group (50 hours) • reading research articles (16 hours) • seminars (6 hours) • literature survey (28 hours) • Individual project (106 hours) • project work (80 hours) • seminars (14 hours) • project presentation and report (12 hours) Tuesday 13–15 W4 Tuesday 15–17 Wednesday 10–12 26/1. Lecture: Introduction. Dependency Grammar 27/1. Lecture: Dependency Parsing. NLP R&D 2/2. Seminar: Research Papers 3/2. Seminar: Research Papers W5 2/2. Seminar: Research Papers W6 Read background literature. Find a project. W7 16/2. Seminar: Project Proposals W8 Project work W9 Project work W10 Project work W11 Project work. Prepare the final report. W12 22/3. Seminar: Final Report W13 Finish the project. Write the final report. 16/2. Seminar: Project Proposals 17/2. Seminar: Project Proposals 8/3. Seminar: Progress Report Project work 22/3. Seminar: Final Report 23/3. Seminar: Final Report Intended learning outcomes By the time you finish the course, you should be able to: • describe and assess the content of scientific articles within natural language processing; reading group • give an account of methods and applications in the area of natural language processing; reading group, project Intended learning outcomes By the time you finish the course, you should be able to: • show an increased ability to independently analyse and solve problems within natural language processing; project • to specify and implement a larger NLP component based on methods and algorithms from the literature; project Intended learning outcomes By the time you finish the course, you should be able to: • apply methods to create NLP data and models, including methods that build on machine learning from text (or speech); project • assess what criteria and measures are relevant for the evaluation of different NLP components and systems and independently carry out an evaluation. project