The class uses point-based assessment of obligatory assignments and requires a minimum sum of 18 points from three assignments. The three assignments will be increasingly rewarding, each worth 8, 10, and 12 points, respectively. In total, we require at least 60 percent of the available points for students to qualify for the final exam. In case one falls below this threshold on one assignment, devoting more effort to the following assignment will make it possible to ‘catch up’ again.
| # | Topic | Publication | Deadline | Points |
| (1) | Bag-of-Words Document Classification | January 24 | February 15 | 8 |
| (2) | Word Embeddings and Semantic Similarity | February 22 | March 8 | 10 |
| (3) | Sentence-Level Sentiment Classification with Convolutional Neural Networks | March 12 | April 5 | 12 |
Because assignments in part build on each other, and because we will be unable to accept late or revised submissions once a ‘model solution’ has been discussed in the laboratory sessions, all submission deadlines need to be firm. We cannot accept submissions after the deadlines. In case of illness, please do get in touch with the instructors as early as possible and in any case prior to the deadline.
We encourage group work for both the three obligatory assignments and the final project. A group is a team of between one and three students who (well before the submission deadline) agree to collaboratively solve the assignment. Once established, the composition of groups can only change inbetween assignments, i.e. after a group has submitted their joint results and before they start to work on a new assignment. Ideally, well-performing groups will start early in the term and stick together for the full semester (and maybe beyond that). Each group member submits the exact same solution, which will be reviewed jointly; each group submission must clearly indicate the group composition, i.e. the full names and UiO account identifiers of all group members.