COMS W4705: Natural Language Processing

[Main] | [General Information] | [Problem Sets]

Instructor: Michael Collins
Time & Location: This semester the class will be taught in a flipped classroom format: details are here.

TAs and Office Hours:

Danning Zheng: Monday 4pm-5.30pm, computer science TA room

Mukund Yelahanka Raghuprasad: Wednesday 11am-12.30pm, computer science TA room

Vu Anh Phung: Thursday 4pm-5.30pm, computer science TA room

Zhuoran Liu: Friday 10am-11.30am, computer science TA room


Past midterms for the class are here: fall 2011, fall 2012, fall 2013, fall 2014, fall 2017, spring 2018.

Lectures: Video lectures are all available by logging in to Courseworks 2.

Date Topics Video Lectures References Flipped Classroom Materials
Week 1 (September 3rd-7th) Introduction to NLP,
Language Modeling (Slides: we will cover slides 1-50 inclusive)
Video lectures in Courseworks: All of Module 1-2; All of Module 3; Sections 4.1, 4.2 in Module 4. Sections 1.1-1.4.1 inclusive of Notes on language modeling (required reading).
Questions (part 1), Solutions (part 1) Questions (part 2), Solutions (part 2)
Week 2 (Sept 10th-14th) Tagging, and Hidden Markov Models (Slides) All videos in Module 6 in courseworks: The tagging problem (10:01) to Summary (1:50) inclusive. Notes on tagging problems, and hidden Markov models (required reading)
Questions, Solutions
Week 3 (Sept 17th-21st) Log-Linear Models (Slides) All videos in Module 15 on Courseworks. Notes on Log-Linear Models (required reading)
Questions, Solutions, Past midterm question
Week 4 (Sept 24th-28th) Parsing, and Context-free Grammars (Slides) Courseworks videos: [1] All of Module 7; [2] Module 8, sections 8-1 to 8-3 inclusive Questions, Solutions
Week 5 (Oct 1st-5th) Probabilistic Context-free Grammars (continued), and lexicalized context-free grammars (Slides part 1) (Slides part 2), (Slides part 3) Courseworks videos: [1] Module 8, sections 8-4 to 8-6; [2] All of Module 9; [3] All of Module 10 Notes on Probabilistic Context-Free Grammars (required reading)
Questions, Solutions
Week 6 (Oct 8th-12th) Log-Linear Models for Tagging, and for history-based parsing (Slides part 1), (Slides part 2). Modules 16 and 17 in courseworks. Notes on MEMMs (Log-Linear Tagging Models) (required reading)
Questions on CRFs, solutions are in section 4 of this note. Additional questions, Solutions
Week 7 (Oct 15th-19th) Feedforward Neural Networks (Slides) Module 22 videos in Courseworks.

Notes on Feedforward Neural Networks (required reading)
Questions, Solutions
Week 8 (Oct 22nd-26th) Mid-term week Note: no flipped classrooms this week.
Week 9 (Oct 29th-Nov 2nd) Computational Graphs, and Backpropagation (Slides) Module 23 videos in Courseworks.

Notes on Computational Graphs, and Backpropagation (required reading)
Questions, Solutions
Week 10 (Nov 5th-9th) Word Embeddings in Feedforward Networks; Tagging and Dependency Parsing using Feedforward Networks (Slides) Module 24 videos in Courseworks.

Week 11 (Nov 12th-16th) Recurrent Networks, and LSTMs, for NLP (Slides) Module 25 in Courseworks. Questions, Solutions
Week 12 (Nov 26th-30th) Recurrent Networks, and attention, for statistical machine translation (Slides) Module 26 in Courseworks. Questions, Solutions