Machine Learning for NLP
Credits: 7,5 hp
Syllabus: 5LN708
Teachers: Christian Hardmeier, Yan Shao, Joakim Nivre
News
 No news is good news.
Schedule
Date  Time  Room  Content  Reading  

1 
27/3 
1416 
Chomsky 
Introduction to machine learning (YS, CH, JN)  
2 
3/4 
1416 
Chomsky 
Decision Trees and Nearest Neighbours. Assignment 1 (CH). 
CiML, Ch. 12 
3 
5/4 
1416 
Chomsky 
Linear Classifiers 1. (CH) 
CiML, Ch. 34 
4 
10/4 
1416 
21024 
Linear Classifiers 2. (CH) 
CiML, Ch. 5 and 7 
5 
17/4 
1416 
Chomsky 
Linear Classifiers from Scratch. Assignment 2. (CH) 
. 
6 
19/4 
1416 
160042 
Linear Classifiers 3. (CH) 
CiML, Ch. 9 
7 
26/4 
1416 
20024 
Generalized linear classifiers. (JN) 
CiML, Ch. 6.2, 17.117.3 
8 
3/5 
1416 
20024 
Introduction to Numpy and TensorFlow. (YS) 

9 
8/5 
1416 
Chomsky 
Linear Classifiers with TensorFlow. Assignment 3. (YS) 

10 
15/5 
1416 
162044 
Neural Networks 1. (YS) 
CiML, Ch. 10, Olah1 
11 
17/5 
1416 
20024 
Neural Networks 2. (YS) 
Olah2 
12 
22/5 
1416 
Chomsky 
Recurrent Neural Networks. Assignment 4. (YS) 

13 
29/5 
1416 
160041 
Machine learning in NLP. (JN) 
Intended Learning Outcomes
In order to pass the course, a student must be able to apply basic principles of machine learning to natural language data,
 apply probability theory and principles of statistical inference to natural language data
 use standard software packages for machine learning,
 implement linear models for classification,
 design simple neural networks for natural language data using some standard library
Examination and Grading Criteria
The course is examined by means of four assignments: Decision trees and nearest neighbor classification. Deadline: April 13.
 Logistic Regression from Scratch. Deadline: April 27.
 Logistic Regression with TensorFlow. Deadline: May 18.
 Recurrent Neural networks. Deadline: June 1.
In order to pass the course, a student must pass all assignments. In order to pass the course with distinction (Väl godkänt), a student must pass at least two assignments with distinction.
Reading List
 CiML = Hal Daumé III. 2011. A Course in Machine Learning. Draft.
 Olah1 Christopher Olah Calculus on Computational Graphs: Backpropagation. Draft.
 Olah2 Christopher Olah Understanding LSTM Networks. Draft.