- 17/03: #4 - MT quiz answers online here
- 17/03: #4 - MT slides online here. Additional reading is also available on the class page.
This year, the course is co-organised by: Chloé Clavel (Inria), Guillaume Wisniewski (LLF, Université Paris Cité), Benoît Sagot (Inria), Rachel Bawden (Inria) and Djamé Seddah (Inria)
For any questions/requests related to this course, please send an email to this address: mva.speech.language@gmail.com
Speech and natural language processing is a subfield of artificial intelligence used in an increasing number of applications. This course will provide an overview and details of techniques and tasks used in the automatic processing of text and speech, covering certain history aspects of the field, the representation of textual and speech data, language modelling, machine translation, sentiment analysis and other labelling tasks, chatbots and speech synthesis and recognition. The aim is to provide the key principles, algorithms and mathematical principles behind the state of the art, and confronting them with the reality of processing real data.
Topics:
- speech features & signal processing
- hidden markov & finite state modeling
- word embeddings
- deep learning for NLP (RNNs, transformers)
- neural language modelling, including large language models (LLMs)
- machine translation
- sentiment analysis
- sequence labelling tasks
- chatbots
- evaluation: comparing human and machine performance
- speech synthesis and speech recognition
Basic linear algebra, calculus, probability theory
The courses consists of 7 three-hours slots. This year, the course will take place in person, distributed over several sites within Paris (please see each separate course for the location).
Each three-hour slot will have a lecture lasting approximately two hours, followed by a quiz and Q&As.
Evaluation consists of 2 parts:
- Quizzes (30% of the total grade): You'll be given a link to an online questionnaire (google form) and will have 30 minutes to complete the questionnaire, which will be activated exactly at 6:00pm and closed down at a time decided on-the-fly by the professors, generally 6:30pm. Any forms submitted after the deadline will be automatically rejected and graded as zero. The quizzes will contain comprehension questions and the best 5 grades out of the 6 quizzes will be used for the average. Between 6:30 and 7:00 there will be a Q&A period where you'll be able to ask questions about the course and quiz.
- Final exam (70% of the total grade): This year (due to time constraints), there will be a final written exam, with theory questions covering topics covered in lectures. The exam will take place on the 29th March 2024 from 4pm-7pm (place TBD).
- #1: Friday 1st March 2024 (4pm-7pm): Introduction (Benoît Sagot), Salle Dussane (45 rue d'Ulm)
- #2: Thursday 7th March 2024 (2pm-5pm): Representation learning and language modelling 1 (Guillaume Wisniewski), Luton Amphitheatre (Faculté de médecine - Cochin campus)
- #3: Friday 8th March 2024 (4pm-7pm): Representation learning and language modelling 2 (Guillaume Wisniewski), Salle Jaurès (29 rue d'Ulm)
- #4: Friday 15th March 2024 (4pm-7pm): Machine translation (Rachel Bawden), Salle Dussane (45 rue d'Ulm)
- #5: Thursday 21st March 2024 (2pm-5pm): Sentiment analysis and labelling tasks (Chloé Clavel), Pavillon 1 (Faculté de médecine - Cordeliers campus)
- #6: Friday 22nd March 2024 (4pm-7pm): Chatbots and real world data challenges (Benoît Sagot and Djamé Seddah), Salle Dussane (45 rue d'Ulm)
- #7: Wednesday 27th March 2024 (9am-12pm): Speech emotion recognition and speech synthesis (Chloé Clavel), Pavillon 3 (Faculté de médecine, Cordeliers campus)
- Exam: Friday 29th March 2024 (4pm-7pm), Salle Dussane (45 rue d'Ulm)
What happens if I get less than 10/20 on average? Can I take another exam?
Any obtained grade is final, so there is no possibility of resitting the exam.
What happens if i cannot be present to the course, and therefore cannot do the online quiz?
Failure to submit the quiz on time will result in a mark of 0/20 for that quiz, unless you can demonstrate that it was materially impossible for you to be present to the course. Such documented requestes should be sent to mva.speech.language@gmail.com together with the name and date of the missed quiz. The best 5 grades out of the 6 quizzes will be used for the average, giving you the possibility of skipping one quiz.