911±¬ÁÏÍø

Machine learning – the ability to learn models to predict and regress from data – has become invaluable tool for robotics and have been applied to almost all areas of robotics, from object recognition to low-level control. The module will cover general concepts, such as regression, classification, density estimation, and dimensionality reduction as well as techniques for computing intractable integrals. Algorithms will be connected to real-world data problems, so that learners can complete research-like tasks, drawing on a range of sources, with a significant level of autonomy. Learners will gain an understanding of the material and main concepts/ theories taught in this module. They will develop skills for analysis and synthesis.

The module covers a range of ethical considerations that arise when machine learning is used in the context of robotics such as bias and fairness, privacy, accountability, safety and ethical decision-making; it also covers sociotechnical topics, such as employment, environment and sustainability and regulation.

Knowledge of research-informed literature will be an outcome of this module. Learners will be able to identify key problem area and choose appropriate methods for their resolution.

Aims:

The aims of this module are to:

• Provide students with a strong foundational understanding of machine learning, particularly for complementary and follow up modules, such object detection and classification and reinforcement learning.
• Provide students with an understanding of the relevance of machine learning specifically with the context of robotics and control.
• Support students with their developing understanding of fundamentals of regression, classification, density estimation, dimensionality reduction, and model selection with the goal of applying this to data-modelling problems.
• Provide an applied context for the use fundamental concepts in object-oriented programming in the creation of programs for machine learning applications.
• Equip students with the knowledge and skills necessary to navigate the ethical complexities of machine learning and contribute to the development of AI systems that align with societal values and norms.

Intended learning outcomes:

1. On successful completion of this module, a student will be able to:
2. Develop a systematic approach to analyzing data using machine learning.
3. Evaluate the quality and suitability of different machine learning methods for modelling data.
4. Examine properties of machine learning algorithms using data interpretation.
5. Develop and build on basic elements of the programming paradigm and the ability to compose these to produce programs that function as intended, scale efficiently in a multi-processor environment, and deliver machine learning results.
6. Analyze the ethical and societal implications of using machine learning in robotics and propose possible solutions to address ethical concerns.

Indicative content:

The following are indicative of the topics the module will typically cover:

• Linear regression.
• Logistic regression and classification.
• Principle component analysis.
• Clustering.
• Introduction to deep learning.
• Feed-forward NNs and ResNets.
• Backprop and autodiff.
• Simple CNNs.

Requisite conditions:

To be eligible to select this module as optional or elective, a student must: be registered on a programme and year of study for which it is formally available.