Computer Science with Artificial Intelligence BSc

• Systems and Devices 3: Advanced Computer Systems

  What you learn (5)

  • explain differences between multiprocessor and single-processor hardware architectures
  • develop software applications designed for multiprocessor architectures
  • optimise software to exploit advanced processor features
  • design memory layouts for data structures in software
  • implement program code to access and manipulate memory-based data structures
  COM00018I 20 cr
• Engineering 1: Systems and Software Engineering

  What you learn (16)

  • explain software engineering terminology and process models
  • apply lifecycle and process models to support a project
  • elicit user and system requirements from stakeholders
  • document user and system requirements clearly
  • design software architecture using object-oriented principles
  • implement object-oriented software from a design
  • estimate project risk to inform decision-making
  • produce realistic project effort and time estimates
  • use continuous integration tools in software development
  • collaborate effectively within a team to complete a project
  • apply software licensing models to project artefacts
  • incorporate third-party software reuse into a project
  • address cyber-security issues throughout the development lifecycle
  • evaluate ethical considerations in software engineering practice
  • write documentation for stages of the engineering lifecycle
  • deliver presentations on project progress and outcomes
  COM00019I 20 cr
• Intelligent Systems: Machine Learning and Optimisation

  What you learn (11)

  • distinguish between strong, weak and general artificial intelligence
  • describe the relationship between computation and artificial intelligence
  • differentiate machine learning from the broader field of artificial intelligence
  • calculate partial derivatives of multivariate functions
  • apply linear regression to model relationships between variables
  • apply logistic regression to classification problems
  • explain the cause and effect of overfitting in machine learning models
  • apply regularisation techniques to prevent overfitting
  • train neural network weights using the backpropagation algorithm
  • analyse ethical arguments concerning artificial intelligence applications
  • evaluate privacy implications arising from machine learning systems
  COM00026I 20 cr
• Theory 3: Computability, Complexity and Logic

  What you learn (14)

  • design a turing machine to recognise a semi-decidable language
  • compute a partial function using a turing machine
  • prove unsolvability of a problem via reduction from a known unsolvable problem
  • prove decidability of a problem by constructing a decision procedure
  • analyse the time complexity of a turing machine
  • analyse the space complexity of a turing machine
  • construct proofs using natural deduction rules
  • determine the truth value of a formula in a given model
  • explain limitations of formal logic
  • explain program behaviour using structural operational semantics
  • reason about program correctness using preconditions and postconditions
  • reason about loop behaviour using invariants in hoare logic
  • distinguish between partial and total correctness of programs
  • apply extended hoare logic laws to prove program termination
  COM00027I 20 cr
• Data: Introduction to Data Science

  What you learn (12)

  • distinguish between data types produced in science, engineering and design contexts
  • apply strategies to ensure the quality of collected data
  • retrieve data from diverse sources and file formats
  • apply inferential statistical methods to test hypotheses about data
  • analyse relationships and features within data sets using statistical procedures
  • select suitable visualisations to present data findings
  • use visual techniques to explore patterns within data sets
  • use relational and non-relational databases to store data
  • construct queries to retrieve data from different database systems
  • identify ethical issues relating to data provenance and individual privacy
  • evaluate the societal impact of data analytics practices
  • apply principles from a professional data protection code of ethics
  COM00028I 20 cr
• Systems and Devices 2: Operating Systems, Security, and Networking

  What you learn (13)

  • write application code using OS-supported concurrency, communication and I/O features
  • explain how hardware features extend basic OS structures introduced previously
  • apply basic OS resource management mechanisms for time and memory allocation
  • apply hardware and OS memory protection mechanisms including memory-mapped I/O
  • apply OS file system mechanisms to enforce information security
  • write concurrent programs at the process level
  • explain how system hardware supports and implements process-level concurrency
  • identify situations of deadlock, livelock and starvation in concurrent systems
  • apply strategies to avoid deadlock, livelock and starvation
  • explain the rationale behind the layered network model
  • develop software that communicates with other systems using OS-level networking sockets
  • describe how networked architectures integrate into an operating system
  • develop simple networked applications using operating system networking knowledge
  COM00029I 20 cr

• Computer Science Project (Artificial Intelligence topic)

  What you learn (15)

  • apply advanced technical skills in a chosen specialist area building on theory
  • design a computer systems engineering solution to a given problem
  • implement a designed computer systems engineering solution
  • evaluate the outcomes of an engineering project against its objectives
  • apply established research methods to generate new knowledge
  • interpret findings produced through recognised research techniques
  • compare alternative approaches to a project problem
  • justify the chosen approach at each project stage
  • identify project tasks feasible within time and resource constraints
  • structure a written project report clearly according to set criteria
  • reference sources to a publishable academic standard
  • critically evaluate the work undertaken and methods applied in the report
  • explain legal, ethical, social and professional issues relevant to the project
  • propose mitigation strategies for identified ethical or professional risks
  • summarise the context, methods and results of a project for a general audience
  COM00015H 40 cr
• Deep Learning

  What you learn (9)

  • explain probability theory as a foundation for machine learning models
  • describe manifold embedding techniques for dimensionality reduction
  • apply kernel methods to transform data representations
  • apply Bayesian regression methods to predict continuous outcomes
  • apply Bayesian classification methods to categorise data
  • apply Bayesian clustering methods to group unlabelled data
  • identify key architectures used in deep learning
  • apply gradient-based optimisation algorithms to train models
  • compare variants of optimisation algorithms used in model training
  COM00049H 20 cr
• Autonomous Robots

  What you learn (7)

  • describe levels of autonomy achievable in robotic systems
  • discuss safety considerations in designing and deploying autonomous robots
  • evaluate ethical implications of autonomous robotic system deployment
  • explain methodologies used to engineer autonomous robotic systems
  • describe how modern robotics middleware is applied in practice
  • build an autonomous robotic system to solve a given task
  • evaluate the strengths and weaknesses of a built robotic solution
  COM00052H 20 cr
• Natural Language Processing

  What you learn (13)

  • describe the process of splitting text into tokens
  • explain how part-of-speech tagging assigns grammatical categories to words
  • explain how parsing determines the syntactic structure of sentences
  • build a text classification pipeline using NLP techniques
  • build a named entity recognition system
  • build a sentiment analysis system using NLP methods
  • apply word embedding techniques to represent language meaning numerically
  • apply n-gram models to estimate word sequence probabilities
  • apply recurrent neural networks to model sequential language data
  • evaluate NLP model outputs using standard performance metrics
  • assess the limitations of NLP models using benchmark datasets
  • describe how classical statistical methods contributed to large language model development
  • describe how neural approaches underpin modern large language models
  COM00069H 20 cr
• Large Language Models

  What you learn (9)

  • critically assess chain-of-thought reasoning methods in modern models
  • describe training, fine-tuning and alignment techniques used to develop language models
  • evaluate explainability approaches and evaluation metrics for language models
  • use modern software frameworks to implement language model applications
  • discuss scalability challenges of deploying large-scale language models
  • measure the energy consumption of training and deploying language models
  • interpret environmental impact data to inform sustainable model development decisions
  • identify ethical concerns relating to bias and fairness in language models
  • address ethical issues of language model use across real-world scenarios
  COM00070H 20 cr
• Engineering LLM-Based Agents and Applications

  What you learn (13)

  • describe common architectural patterns for combining LLMs with planning and memory modules
  • compare design patterns integrating retrieval and external tool use in LLM agents
  • build an intelligent software agent using LLM components
  • deploy a functioning agent system to a working environment
  • design an autonomous workflow combining LLM reasoning with retrieval mechanisms
  • construct multi-step action plans that link reasoning to execution
  • implement a practical agent application such as a task assistant or planner
  • integrate external APIs and services into an agent application
  • evaluate agent performance using qualitative and quantitative measures
  • analyse limitations of agent systems relating to reliability and latency
  • identify safety and ethical risks arising in agent design
  • propose mitigations for hallucination and over-reliance on automated agents
  • explain how systems thinking supports evaluation of LLM performance in practice
  COM00071H 20 cr
• AI Search and Logic

  What you learn (13)

  • formulate a search problem using states, actions and goal conditions
  • identify a suitable heuristic for a given search problem
  • translate a scenario into propositional logic statements
  • perform logical inference on a propositional representation using a sat solver
  • model a discrete optimisation problem using constraint programming
  • solve a discrete optimisation problem with a constraint programming system
  • justify the choice of a search algorithm over alternatives
  • apply a chosen search algorithm to solve a problem
  • select an adversarial game-tree search method for a given game
  • design a heuristic to support adversarial search when needed
  • apply an adversarial search method to solve a game
  • describe common algorithms used in sat, local search and constraint solvers
  • apply solver algorithms to small example problems
  COM00191M 20 cr

Question texts are published by the Office for Students; the dataset carries only the question numbers.