Machine Learning

Learning outcome 1: Explain, evaluate and apply Machine Learning.

Learning outcome 2: State, select, apply and evaluate Machine Learning approaches.

Learning outcome 3: Explain, evaluate and apply a range of Machine Learning algorithms.

Learning outcome 4: Critically evaluate an application of ML algorithms to a given problem.

Learning outcome 5: Synthesis, analyse and evaluation literature relating to a given topic in verbal and written format.

The group will facilitate discussions, identifying the domain problem boundaries and shared resources, e.g. datasets, reviewing the domain problem general literature, and facilitate comparisons between solutions to appreciate the multi-disciplinary nature of AI and range of techniques.

Each individual, however, will explore a topic and/or a solution, e.g. a specific technique, apply it to the domain problem using appropriate tools, e.g. programming or a platform. Each individual will review the literature specific to their topic/solution/technique and demonstrate its working, e.g. testing.

This coursework is an exploration of AI and its applications in an AI-enabled world. Thus, it will primarily consist of literature appraisal and practical exploration of AI algorithms. You are not expected to write the code of these algorithms from scratch, instead you can explore and evaluate the algorithms using appropriate tools over examples or case studies from your chosen domain.

Task 1. Deciding on the Problem Domain to be explored. Please consider the following:

a. Form a group of 2-4 members.
b. Decide on your chosen domain from the following themes:
i. Digital Health, e.g. health informatics, drug discovery, health analytics, etc.
ii. Smart Environments, e.g. smart cities, smart vehicles, smart manufacturing, smart education, etc.
iii. AI Infrastructures, e.g. Internet of Thinks (IoT), robotics, big data storage centres management.
iv. Trustworthy AI: secure AI, AI for data security and privacy, human-centred AI, Affective AI.
c. Discuss in the group the challenges within your chosen domain and the Machine Learning approaches that can be applicable and how these approaches may help to address these challenges, leading to specifying an appropriate Machine Learning application within your chosen domain.

Task 2. Once you have identified the application domain and the challenges, evaluate the Machine Learning approaches and the techniques provided by these approaches that can be used to address these challenges leading to the identification of specific techniques and algorithms to be trialled. Please consider the following:

a. Identify appropriate tools to be used, e.g. Python libraries, MATLAB toolbox, Open Source platforms and simulators, such as WEKA, RapidMiner, NetLogo, and Jason.
b. Identify appropriate examples to demonstrate the applicability of the chosen Machine
Learning algorithms, e.g. relevant datasets, etc.

Task 3. Each individual in the team is to choose no more than 2 contrasting Machine Learning algorithms and explore their applicability to the chosen application domain in answering one identified challenge within that domain.

a. Explain the algorithm(s) working mechanisms. You can use pseudocode or diagrams.
b. Apply the algorithms(s) to some of the identified examples or case studies using appropriate tool or simulator, e.g. WEKA, RapidMiner, SAS, Matlab, etc.
c. Evaluate the performance of the algorithm(s).

Task 4. Each individual in the team is to summarise their work from Task 3 in an executive summary of no more than 2 pages in IEEE format.

a. Synthesis, analyse and evaluate the literature relating to your chosen domain and challenges, i.e. brief summary of Task 1 and Task 2
b. Critically evaluate in depth your application of your chosen Machine Learning algorithm(s) to your chosen application domain, i.e. concise and in-depth summary of Task 3 complemented by critical evaluation of the results or conclusions reached.

Attachment:- Machine Learning.rar