Elective in Artificial Intelligence
A.Y. 2023/2024
Description of the course
The course gives 12 credits within the
Master in Artificial Intelligence and Robotics
It is structured in the four modules described below. Please refer to each module, following the links in this page for additional details.
Professor in charge of exam registration is Prof. Fabio Patrizi.
Any two modules from Electives in AI can be taken to cover Research Topics in AI (6CFU), provided Electives in AI has not been inclued in the study plan.
Lectures
All the lectures will be given in the second semester.
Additional information are provided in the web sites of the modules.
Exams
Each module has specific modalities. Please check the module web sites.
Exam dates: the following dates represent deadlines for evaluation of projects and registration of the exams:
June 15
July 15
September 15
October 31
January 15
February 15
March 31
Projects delivered by any of these dates will be evaluated within 10 days.
For registration of Elective in AI exam, after completing all the modules, send an e-mail to the Professor in charge (patrizi@diag.uniroma1.it with subject of email starting with [EAI]) providing the following information: name, matricola, list of given exams (title of the module, professor of the module, and grade).
Registration will be completed within 15 days of the above deadlines, provided all the modules have been completed (i.e., the vote is agreed with each Professor) and the email requesting for registration arrives at least 2 days before the end of the month (i.e., 13 days after the above deadlines).
Total grade for 12 CFU is computed as follows:
- 30 with Laude in each module is counted as 31
- all the votes are weighted averaged
- the final grade is rounded to the closest integer (0.5 is rounded to the highest value, e.g. 29.5 -> 30)
- Laude is granted with weighted average >= 30.5 (e.g., 30 L x 6 CFU + 30 x 6 CFU -> 30 L)
AI for Visual Perception in HCI & HRI
Prof. Christian Napoli
3 CFU - 1st Semester
As computer science, assistive technologies and robotics evolve towards application fields in which humans cooperate with machines, working closer and closer, the requirements for human computer interactions increase. Visual perception is an important component for human–machine interaction processes in all kinds of computer systems. Interaction between humans and computers depends on the reliability of the perception systems, and, above all, the vision system. The analysis of activities, motions, skills, and behaviors of humans and robots are generally addressed by using the features of a moving human body (or body part). The human motion behavior is then analyzed by body movement kinematics, and the trajectory of the target is used to identify the objects and the human target. The process of human target identification and gesture recognition in a quite non-trivial problem. In this series of lectures we will focus on the context of Human-Robot Interaction (HRI) along with the related problems on the field of vision and perception, applied to robotic systems. We will devise the typical characteristics of vision and perception related hardware device, as well as the relative software systems and solutions. We will explore the known approaches characterizing well known visual recognition systems, as well as the most important algorithmic solutions for people targeting and body parts recognition. A theoretical and practical framework will be given with several example. Finally we will discuss the state of the art on human-centric vision analysis and explain the importance of the matter relatively to human-based interfaces of computer/robots with special interest in human motion and activity recognition. We will also devise several tracking systems and motion oriented context and object recognition techniques, with emphasis on deep learning techniques applied to visual recognition. Finally we will compare the applicability of such techniques to human motion classification and the related application on the field of Human-Computer Interaction.
Hot Topics in Natural Language Processing (HoTNLP)
where: Room B2 - Via Ariosto 25
when: Monday, Tuesday, Wednesday (8.30-10.00), starting September 25th
3 CFU - 1st Semester
The field of Natural Language Processing (NLP) has witnessed unprecedented growth and innovation in recent years, making it one of the most dynamic and exciting domains in artificial intelligence. This course will explore the most current and groundbreaking developments in NLP, offering students a deep dive into the latest advancements, techniques, and challenges that define the field today, including:
Transformer Models and Beyond: understanding the impact of models like BERT, GPT-3, and their successors
Low-Resource and Cross-Lingual NLP: exploring strategies for NLP in languages with limited resources and in cross-lingual applications
Multimodal NLP: investigating the integration of text with other modalities such as images and audio
Event extraction: automatically identifying and extracting structured information about events, such as dates, locations, participants, and their roles, from unstructured text data
Dialogue: exploring some challenges of conversational AI.
Commonsense knowledge and reasoning: addressing the challenges of identifying and encoding commonsense information and reasoning with it
Summarization: introducing abstractive and extractive methods, and their applications in distilling information from vast textual datasets
Interpretable NLP: sketching some ideas behind making NLP models interpretable and transparent, and its importance
3 CFU - 2nd Semester
One of the most important goals of Artificial Intelligence concerns the development of intelligent agents, such as robots, that are able to deliberate their course of actions. This course introduces several approaches to realize this goal, with a particular focus on advanced forms of automated behaviour synthesis, based on temporal logics and action theories, in both deterministic and non-deterministic settings.
Students are encouraged to realize projects which integrate such high-level forms of reasoning, with state-of-the-art probabilistic robotics, computer vision, and machine learning tools.
Prof. Luca Iocchi
3 CFU - 2nd Semester
Applications involving robots interacting with people are gaining increasing interest. The course will provide an overview of recent methods and techniques for Human-Robot Interaction and Social Robotics. Several interaction modalities will be discussed in the course, including: vision, speech, body motion, user interfaces, etc.