Forskjeller

Her vises forskjeller mellom den valgte versjonen og den nåværende versjonen av dokumentet.

Lenk til denne sammenligningen

Begge sider forrige revisjonForrige revisjon
Neste revisjon		Forrige revisjon
Siste revisjonBegge sider neste revisjon
emner:fordypning:ttk23 [2019/10/08 20:52] – [Exam] scibiliaemner:fordypning:ttk23 [2020/10/24 14:39] – [Guest lectures] scibilia
Linje 3: Linje 3:
 **Instructors:** Anastasios Lekkas (NTNU) and Francesco Scibilia (Equinor/NTNU)  **Instructors:** Anastasios Lekkas (NTNU) and Francesco Scibilia (Equinor/NTNU) 
  
-**Updated Info:** The course starts on __9 October 2019__, 10:15-13:00, Room R52 (previously R53). Note: Lecture 7 is in room S21.+**Updated Info:** The first lecture will take place on Monday 5 October between 12:15-14:00 at S4 (Sentralbygg 1).  
 + 
 +**Prerequisites:** For the academic part, it's useful to have some past knowledge on how to train neural networks. 
 ===== Description ===== ===== Description =====
  
Linje 12: Linje 14:
  
 ==== Theory lectures (by Anastasios Lekkas) ==== ==== Theory lectures (by Anastasios Lekkas) ====
 +The lecture plan below is from last year. For the 2020 lectures, it will be enhanced with additional algorithms and details regarding their implementation.
  
-**Lecture 1** **(09/10/19, 10:15-13:00, R52):** +**Lecture 1 (October 5, 2020, 12:15 - 14:00, S4 (Sentralbygg 1))**  
   * Course introduction: Objectives, overview of lectures, assignments, exam info.   * Course introduction: Objectives, overview of lectures, assignments, exam info.
   * Autonomy in the scope of the course. An introduction to reinforcement learning and its connection with dynamic programming.    * Autonomy in the scope of the course. An introduction to reinforcement learning and its connection with dynamic programming. 
   * Markov Decision Processes and problem formulation   * Markov Decision Processes and problem formulation
  
-**Lecture 2** **(16/10/19, 10:15-13:00, R52):** +**Lecture 2 (October 12, 2020, 12:15 - 14:00, S4 (Sentralbygg 1))** 
   * Optimal policies in discrete environments (Part 1): Value iteration and policy iteration   * Optimal policies in discrete environments (Part 1): Value iteration and policy iteration
   * Optimal policies in discrete environments (Part 2): Q-learning and SARSA   * Optimal policies in discrete environments (Part 2): Q-learning and SARSA
  
-**Lecture 3** **(23/10/19, 10:15-13:00, R52):** +**Lecture 3 (October 19, 2020, 12:15 - 14:00, S4 (Sentralbygg 1))**
   * Deep learning and deep reinforcement learning.    * Deep learning and deep reinforcement learning. 
   * The DQN algorithm.   * The DQN algorithm.
   * Introduction to policy gradient algorithms for continuous action and state spaces. REINFORCE algorithm   * Introduction to policy gradient algorithms for continuous action and state spaces. REINFORCE algorithm
  
-**Lecture 4** **(30/10/19, 10:15-12:00, R52):** +**Lecture 4 (October 26, 2020, 12:15 - 14:00, S4 (Sentralbygg 1))**  
   * Deep deterministic policy gradients (DDPG)   * Deep deterministic policy gradients (DDPG)
-  * Model predictive control and reinforcement learning (by Prof. Sebastien Gros)+  * Proximal policy optimization (PPO)
  
 ==== Industry lectures (by Francesco Scibilia) ==== ==== Industry lectures (by Francesco Scibilia) ====
  
-**Lecture 5** **(06/11/19, 10:15-12:00, R52):** Artificial intelligence in autonomous robotics systems: what is an actionable definition in an industrial setting. Different levels of autonomy. Hierarchical architecture (Sense‐plan‐act and behaviorbased substrates) and autonomy layers. A control theory approach to planning and action – Model Predictive Control.+**Lecture 5 (November 2, 2020, 12:15 - 14:00, S4 (Sentralbygg 1))**   
 +  
 +Artificial intelligence in autonomous robotics systems: what is an actionable definition in an industrial setting. Different levels of autonomy. Hierarchical architecture (Sense‐plan‐act and behaviorbased substrates) and autonomy layers. Data and connectivity aspects for autonomy.
  
-**Lecture 6** **(13/11/19, 10:15-12:00, R52):** Autonomous mobile robotics systems – environment perception vs mission sensors, localization and mapping, navigation and sense & avoid. Considerations on implementing autonomous systems in industrial applications as: system integration, ethics concerns, solutions fit to existing organization and systems, data management and intellectual property aspects.+**Lecture 6 (November 9, 2020, 12:15 - 14:00, TBD)**  
 +  
 +AI Robotics, market value chain considerations. Operational considerations on implementing autonomous systems in industrial applications as: business models, system integration, solutions fit to existing customer infrastructure and systems, emerging industrial information standards aspects.
  
  
 ==== Guest lectures ==== ==== Guest lectures ====
-**Lecture 7** **(18/11/19, 10:15-12:00, S21):** 
-More info to be announced soon. 
  
-==== Exam ==== +**Lecture 7 (November 6, 2020, 11:45 - 14:00, online(details to come) )**   
-**28 & 29 November**+  
 +  * 11:45-12:00 Connecting and Intro – Francesco  
 +  * 12:00-12:50 Guest lecture + QA – Steffan Sørenes, Leading Advisor IT Architecture at Equinor 
 +  * 12:50-13:10 Break 
 +  * 13:10-14:00 Guest lecture + QA – Fakhri Landolsi, Manager Data Science at Equinor 
 + 
 +==== Oral Exam ==== 
 +November 25-26.
2020/10/24 14:40, scibilia