Introduction to Deep Learning
Winter School at Universitat Politècnica de Catalunya (2018)
Deep learning technologies are at the core of the current revolution in artificial intelligence for multimedia data analysis. The convergence of large-scale annotated datasets and affordable GPU hardware has allowed the training of neural networks for data analysis tasks which were previously addressed with hand-crafted features. Architectures such as convolutional neural networks, recurrent neural networks or Q-nets for reinforcement learning have shaped a brand new scenario in signal processing. This course will cover the basic principles of deep learning from both an algorithmic and computational perspectives.
Lectures (30%) - Room: D5-010
Course will be divided in modules of half an hour covering the following topics:
- 22/01 10:00 D1L1 (XG) Welcome
- 22/01 10:30 D1L2 (VV) Machine Learning
- 22/01 11:00 D1L3 (AB) Perceptron
- 22/01 11:30 D1L4 (ES) Multi-Layer Perceptron
- 23/01 10:00 D2L1 (ES) Backpropagation
- 23/01 10:30 D2L2 (VV) Optimizers
- 23/01 11:00 D2L3 (JR) Loss functions
- 23/01 11:30 D2L4 (JR) Methodology
- 24/01 10:00 D3L1 (VV) Convolutional Neural Networks - CNNs
- 24/01 10:30 D3L2 (RM) Transfer learning
- 24/01 11:00 D3L3 (MRC) Recurrent Neural Networks - RNNs
- 24/01 11:30 D3L4 (MRC) Gated RNNs
- 24/01 10:00 D4L1 (MRC) Attention Models
- 25/01 10:30 D4L2 (XG) Unsupervised learning
- 25/01 11:00 D4L3 (VV) Adversarial Training
- 25/01 11:30 D4L4 (XG) Architectures
- 30/01 11:30 Guest: Joost van de Weijer (Computer Vision Center) - RSVP here
- 30/01 12:15 Guest: Joan Serrà (Telefónica)- RSVP here
Labs (30%) - Room: D5-004
The course will contain guided hands on lab provided by the NVIDIA Deep Learning Institute.
- 22/01 Lab1: Image Classification with DIGITS
- 23/01 Lab2: Linear Classification with Tensorflow
- 24/01 Lab3: Modeling Time Series Data with Recurrent Neural Networks in Keras
- 25/01 Lab4: Identifying Whale Sounds with Audio Classification
Project (40%) - Room: D5-004
Students will work in teams to develop a machine learning research project that will be presented in an oral presentation during the final day of the course.
- Study Programs: Bachelor degrees at at ETSETB TelecomBCN and [Centre de Formació Interdisciplinària Superior (CFIS)(http://cfis.upc.edu/ca) from the Universitat Politecnica de Catalunya.
- Course code and official guide: 230235 - IDL
- ECTS credits: 2 ECTS
- Teaching language: English
- Semester: Autumn 2017
- Class Schedule: 22, 23, 24, 25 & 30 January 2018 (10am-12pm Lectures, 12pm-2pm Lab)
- Capacity: 40 BSc students
- Location: Campus Nord UPC, Module D5, Room 010
- Course on Piazza
Registration procedure depends on the student profile:
- Bachelor students at ETSETB who can register the 2 ECTS: Follow the regular schedule from your academic office. There is an extraordinary registration period between 11 and 14 January 2018.
- Bachelor students at ETSTEB who cannot register the 2 ECTS: Fill in this form.
- Master students at ETSETB: This Winter School is a light version of the full MSc course of Deep learning for Artificial Intelligence, taught during the Autumn semester. So we encourage you to sign up for this other one. However, if you would to register to the Winter School on Deep Learning for Speech and Language (DLSL) or the Summer School on Deep Learning for Computer Vision (DLCV), you can still follow this introductory course in the morning and sign up officially for DLSL in the afternoons and for DLCV in July 2018. If this is your case, fill in this form so we can have a seat for you to the introductory one, and register to DLSL.
- Students at UPC but not in ETSETB: Contact the your academic office and request being allowed to take this course. If accepted, contact ETSETB academic office and request more details.
- Mobility students: If your host institution has signed an agreement with UPC ETSETB Telecom BCN school, you can request a mobility from your host institution and sign up for the course under the same conditions as ETSETB students.
- Industry members and any other profile: You must apply for being accepted in the course and cover the 100% cost of the ECTS credits, without the support of the public funds. This corresponds to 143,08 € per ECTS credit (Summer 2016). If you are interested in this option, please contact the ETSETB Telecom BCN academic office.
- Deep Learning for Artificial Intelligence. UPC TelecomBCN 2017.
- Deep Learning for Computer Vision UPC TelecomBCN: 2016 2017
- Deep Learning for Speech and Language. UPC TelecomBCN 2017.
- Deep Learning for Multimedia. Insight Dublin City University 2017.
- Jordi Torres, Intro Labs to Keras, PyTorch, TensorFlow and Google Cloud. UPC TelecomBCN 2017.
- Amaia Salvador and Santiago Pascual. “Hands on Keras and TensorFlow”. Persontyle 2017.
- Xavier Giro-i-Nieto, “Deep learning for computer vision: Image, Object, Videos Analytics and Beyond”. LaSalle URL. May 2016.