Educational objectives General objectives:
The course is intended as a broad overview to neural networks, as used today in a number of applicative fields. It provides a strong theoretical and practical understanding of how neural networks and modern deep networks are designed and implemented, highlighting the most common components, ideas, and current limitations.
Specific objectives:
From a theoretical point of view, we will review the general paradigm of building differentiable models that can be optimized end-to-end with gradient descent from data. We will then overview essential components to design architectures able to work on images (convolutive layers), sequences (recurrent layers), and sets (transformer layers). The last part of the course will then focus on a selection of important research topics, including graph neural networks, continual learning, and generative models.
Knowledge and understanding:
At the end of the course, the student will have a broad understanding of how deep networks work in practice, with the capability of implementing new components from scratch, re-using existing models, or designing new architectures for problems beyond the overview of the course.
Critical and judgment skills:
The student is expected to be able to analyze a new problem requiring machine learning, and design the appropriate neural network based solution to tackle it, understanding both its strengths and its drawbacks.
Communication skills:
The course will foster communication skills in terms of being able to describe (in both a technical and non-technical way) the mathematics underlying the models, as long as writing clear and understandable code for its implementation.
Learning ability:
Beyond the topics of the course, the student will be able to autonomously study new topics on the research frontier, and navigate the current scientific literature and software panorama.
|
Educational objectives General Objectives.
The Reinforcement Learning (RL) course aims to introduce students to fundamental and advanced techniques of RL, a significant area within artificial intelligence and machine learning. Students will gain skills to design and implement algorithms that enable systems to learn and improve autonomously through experience, optimizing their decisions in real-time.
Specific Objectives.
Students will explore key concepts of RL such as decision policies, Markov Decision Processes, Q-learning, and deep reinforcement learning. They will learn to:
Model complex problems using the RL approach.
Develop and implement algorithms like Q-learning and Deep Q-Networks (DQN).
Apply RL techniques in real-world scenarios like robotics, gaming, etc.
Knowledge and Understanding:
In-depth knowledge of basic and advanced RL algorithms.
Understanding of reward-based learning models and their practical applications.
Ability to interpret the results of RL algorithms and evaluate their effectiveness in various contexts.
Applying Knowledge and Understanding:
Use software frameworks like TensorFlow or PyTorch to implement and test RL algorithms.
Analyze current research case studies and projects to understand real-world RL applications.
Develop functional prototypes using RL to solve specific problems.
Autonomy of Judgment:
Students will develop the ability to critically assess RL algorithms, considering their applicability, efficiency, and potential biases. They will also be able to select the most appropriate algorithm for a given problem.
Communication Skills:
Students will learn to effectively communicate RL concepts, algorithm design decisions, and outcomes to both technical and non-technical audiences using a variety of communication media.
Next Study Abilities:
This course will prepare students to pursue advanced studies and research in RL, providing the necessary foundation to tackle open problems and innovate in the field. Students will be encouraged to actively contribute to the scientific community through publications, conferences, and collaborations.
|
Educational objectives General Objectives:
Acquiring knowledge on the basic tools for probabilistic state estimation in robotics.
Being able to apply these tools to real study cases and to implement working solutions.
Evaluate the quality of a state estimator.
Specific Objectives:
Knowledge and Understanding:
- how to manipulate probability distributions, in particular Gaussians
- the basics of filtering (hisrogram filters, Gaussian filters, particle filters)
- the generic model for a stationary non-linear or linear
- Dense and Sparse formulation of minimization algorithms (Gauss-Newton, Levenberg Marquardt)
- The problem of Data Association, and typical tools to approach it (RANSAC, Heuristics)
- Typical study cases of estimation problems in robotics (Calibration, Localization, Mapping and SLAM)
Applying Knowledge and Understanding:
- Being able to model a problem and to adapt the tools to its solution.
- Develop a functioning estimator.
Making Judgements:
- Being able to analyze the pros and contra of different solutions to the same problem.
- Spot the tools applicable to solve all subtasks in the design of an estimator.
These abilities are supported by the Project to be developed as a part of the exam.
The course interleaves theory and practice. During the practicals the students are asked to
complete code snippets provided by the teacher and to run their programs on real study cases.
Communication Skills:
- Acquire a common language to describe estimators and a development methodology
that supports interaction between developers by defining a standard set of goals.
Learning Skills:
The student will possess the abilities and the skills to approach general estimation problems.
The examples in the domain of navigation provided during the course serve as study cases.
The indivudal topics learned (Gaussian Manipulation, Filtering Designs, Minimization)
are useful instruments to approach a far more general class of problems
|
Educational objectives General objectives:
The course is intended as a broad overview to neural networks, as used today in a number of applicative fields. It provides a strong theoretical and practical understanding of how neural networks and modern deep networks are designed and implemented, highlighting the most common components, ideas, and current limitations.
Specific objectives:
From a theoretical point of view, we will review the general paradigm of building differentiable models that can be optimized end-to-end with gradient descent from data. We will then overview essential components to design architectures able to work on images (convolutive layers), sequences (recurrent layers), and sets (transformer layers). The last part of the course will then focus on a selection of important research topics, including graph neural networks, continual learning, and generative models.
Knowledge and understanding:
At the end of the course, the student will have a broad understanding of how deep networks work in practice, with the capability of implementing new components from scratch, re-using existing models, or designing new architectures for problems beyond the overview of the course.
Critical and judgment skills:
The student is expected to be able to analyze a new problem requiring machine learning, and design the appropriate neural network based solution to tackle it, understanding both its strengths and its drawbacks.
Communication skills:
The course will foster communication skills in terms of being able to describe (in both a technical and non-technical way) the mathematics underlying the models, as long as writing clear and understandable code for its implementation.
Learning ability:
Beyond the topics of the course, the student will be able to autonomously study new topics on the research frontier, and navigate the current scientific literature and software panorama.
|
Educational objectives Introduction to the basic robotic technologies in the medical context, with particular emphasis on surgical robotics.
Expected learning results: Knowledge of the main robotic surgical systems, of the challenges and methodologies of medical robot design and control.
Expected competence in:
- critically reading a scientific paper describing medical robotics technologies;
- discussing in detail the state of the art of robotic applications in medicine;
- estimating potential benefits deriving from the introduction of robotic technologies in a medical procedure;
- arguing the development of a particular technology not yet available or experimentally validated;
- communicating and collaborating with people with different technical background;
- evaluating clinical, social and economical constraints in implementing a robotic technology in a medical context;
- design control scheme for teleoperation of medical robots and for shared execution of surgical tasks between humans and robots.
|
Educational objectives * General objectives
The course aims to introduce the principles, methodologies, and applications of the main engineering techniques used to study and interact with neural systems.
* Specific objectives
- Knowledge and understanding
Students will learn the basics of the human brain functioning and organization at different scales, and to the main applications of engineering and information technologies to neuroscience
- Applying knowledge and understanding
Students will familiarize with basic tools to utilize to acquire, process and decode neurophysiological and muscular signals and to interface them with artificial devices
- Critical and judgment skills
Students will learn how to choose the most suitable control methodology for a specific problem and to evaluate the complexity of the proposed solution.
- Communication skills
Students will learn to communicate in a multidisciplinary context the main issues of interfacing neurophysiological signals with artificial systems, and to convey possible design choices for this purpose.
- Learning ability
Students will develop a mindset oriented to independent learning of advanced concepts not covered in the course.
|
Educational objectives General Objectives
The goal of the course is to provide an overview of state-of-the-art natural language processing techniques and their applications.
Specific Objectives
Students will learn the principles of automatic language processing, understanding how machines can interpret, generate and respond to human language. This includes topics such as word representation, word and sense embeddings, neural architectures for NLP, machine translation, and more general text generation.
Knowledge and Understanding
-) Knowledge of neural network architectures, such as recurrent neural networks and Transformers, used for natural language processing.
-) Knowledge of supervised and unsupervised learning methods in NLP.-) Knowledge of lexical and phrasal computational semantics techniques.
-) Understanding of language models for interpreting and generating text.
Applying knowledge and understanding:
-) How to develop models for understanding language
-) How to develop models for generating language
-) How to use neural architectures for NLPAutonomy of Judgment.
Autonomy of Judgment
Students will be able to evaluate the effectiveness of NLP techniques in different applications.
Communication Skills
Students will be able to explain the principles and techniques of natural language processing.
Next Study Abilities
Students interested in research will discover what are the main open challenges in the area of NLP, obtaining the necessary foundation for more in-depth studies in the field.
|
