Quantum Computing Hackathon

Natwest / ETH / Warwick / Surrey / NQCC / UCL

Theme: Quantum Portfolio Optimisation

• Tackles the classic “mean–variance” portfolio problem: choosing asset weights that trade off risk and return.
• Translates the problem into a QUBO form so it can be solved with the Quantum Approximate Optimisation Algorithm (QAOA).
• Targets IBM’s 127-qubit Brisbane device, carefully designing circuits within depth and connectivity limits and adding error-mitigation.
• Uses modern optimisation techniques (Proximal Policy Optimisation and Bayesian optimisation) to automatically tune QAOA parameters.
• Compares quantum-generated portfolios with classical benchmarks to see when quantum methods give better, or more robust, allocations.

Ecole Normale Supérieure - PSL / Ecole Polytechnique

Theme: Quantum Portfolio Optimisation

• Builds a full “quantum pipeline” to choose and weight stocks in an investment portfolio.
• First uses quantum kernels plus machine-learning to filter thousands of stocks down to the most promising few.
• A neural network scores these assets, then prepares a tailored quantum starting state to speed up the search for good portfolios.
• Advanced quantum optimisation (QAOA variants) is used to balance return, risk and even ESG scores when constructing portfolios.
• Noise-mitigation tools and hardware tests on IBM devices explore how far today’s quantum machines can be pushed for real finance tasks.

Quantum Buddies

Sheffield University / Leeds University / Imperial College London

Theme: Market Simulation with Quantum

• Aims to forecast market price movements by learning the full probability distribution of financial time series, not just a single prediction.
• Uses a Quantum Circuit Born Machine (a quantum generative model) to mimic how market prices behave, based on historical data.
• Explores different circuit designs and feature mappings so that financial signals are encoded efficiently into qubits.
• Combines this with Quantum Markov Chain Monte Carlo to sample many realistic future scenarios and quantify uncertainty.
• Tests on quantum hardware look at whether these hybrid methods can capture complex market dynamics better than classical models.

Touraregs

Theme: Post Trade Innovation

• Seeks to unlock value from unstructured financial text, such as KPI statements and client metadata.
• Uses quantum natural language processing to classify statements and potentially flag patterns linked to collusion or misconduct.
• Represents each sentence as a quantum circuit where words control the gate parameters, creating a “quantum fingerprint” of meaning.
• Builds a quantum kernel that measures similarity between sentences, using the lambeq toolkit and a compute–uncompute design.
• The goal is a scalable, quantum-enhanced text analysis tool that can handle linguistic ambiguity better than standard methods.