Week 7: Lexical semantics and word embeddings

Reminders and announcements

• Solutions to Tutorial 2 are now available.
• Lab 3 worksheet is now available, for the labs this week on text classification.
  • You should be able to complete the lab during your scheduled timeslot, but you may wish to review the material on Naive Bayes and multinomial logistic regression (Week 2) before attending.
• Assignment 2 partnering form, due by noon on Thu if you want to work with a partner.
  • You may not work with the same partner as you did for Assignment 1.
  • If you found your own partner, one of you should fill in the form to tell us who the partner is.
  • If you want us to assign you a partner, fill in the form to tell us that.
  • Unlike for Assignment 1, we will only assign you a partner if you fill in the form. If you want to work alone, you don't need to fill in the form. However, previous experience suggests that pairs tend to do better on the assignment than individuals.
• Tutorial exercises for the Week 8 groups are here. The questions are based on CKY parsing (W5 L2) and issues with web data (W6 L3 reading). As usual you will need to spend some time on the reading and exercises in advance.

Overview of this week

Welcome to Week 7 from Edoardo, I am back in the lecture theatre this week. During this week's lectures, we will address the question: how can we represent the meaning of a word in a way that is machine-readable? We will start from the theoretical foundations of lexical semantics, including relational and distributional definitions of meaning. Then, we will cover two families of word representations: sparse and dense word vectors, highlighting their pros and cons and their role in current machine learning models.

Lectures and reading