Qronon

Members

First name (team leader)
Osama
Last name
Ahmed
Organisation name
Imperial College London
Organisation type
Academic (Student)
Organisation location
United Kingdom
First name
Sallar
Last name
Ali Qazi
Organisation name
Imperial College London
Organisation type
Academic (Student)
Organisation location
United Kingdom
First name
Luca
Last name
Magri
Organisation name
Imperial College London
Organisation type
Research Organisation (Academic, Independent, etc.)
Organisation location
United Kingdom

Models

Model name

QRCML
Number of individuals supporting model development:
1-5
Maximum number of Central Processing Units (CPUs) supporting model development or forecast production:
8-48
Maximum number of Graphics Processing Units (GPUs) supporting model development or forecast production:
< 4
How would you best classify the IT system used for model development or forecast production:
Single node system

Model summary questionnaire for model QRCML

Please note that the list below shows all questionnaires submitted for this model.
They are displayed from the most recent to the earliest, covering each 13-week competition period in which the team competed with this model.

Which of the following descriptions best represent the overarching design of your forecasting model?
  • Machine learning-based weather prediction.
  • Ensemble-based model, aggregating multiple predictions to assess uncertainty and variability.
  • Other: A hybrid quantum-classical machine learning model trained and used for inference.
What techniques did you use to initialise your model? (For example: data sources and processing of initial conditions)
ERA5 data sets are used for initialization, given that the data is only available 6 days after. Our model is actually initialized 10 days before the actual competition start date, compared to other models.
If any, what data does your model rely on for real-time forecasting purposes?
ERA-5 and the data set provided by the competition function.
What types of datasets were used for model training? (For example: observational datasets, reanalysis data, NWP outputs or satellite data)
ERA5 Reanalysis data from the 1st January 1979 to 31st December 2024 were used. The data was downloaded at six-hourly intervals (0, 6, 12, and 18 UTC) at a 1.5-degree resolution. Seven-day rolling moving averages were computed for temperature and pressure.
Please provide an overview of your final ML/AI model architecture (For example: key design features, specific algorithms or frameworks used, and any pre- or post-processing steps)
We use a Quantum Machine Learning model trained on a single GPU and a small workstation. Our novel architecture makes it lighter to train and improve the forecasting period with fewer resources. As compared to other big models, our QML model is faster to train, taking only a few minutes to train for each forecasting period and a few seconds of inference cost per ensemble.
Have you published or presented any work related to this forecasting model? If yes, could you share references or links?
Some foundational work of our model is open-sourced and published in 3 peer-reviewed journals, including the (i) American Physical Society, Physical Review Research - https://doi.org/10.1103/PhysRevResearch.6.043082 (ii) Springer Nature, Quantum Machine Intelligence - https://doi.org/10.1007/s42484-025-00261-9 (iii) The Royal Society, Proceedings of the Royal Society A - https://doi.org/10.1098/rspa.2025.0550 We should note that our current employed model is massively improved from this published work, and we are looking to improve and scale the architecture further by the end of this competition.
Before submitting your forecasts to the AI Weather Quest, did you validate your model against observational or independent datasets? If so, how?
We validated on observational data sets and computed the RPSS of the August forecast by initializing the model in July.
Did you face any challenges during model development, and how did you address them?
We scaled our architecture as compared to previously published work. The new architecture will be published soon in a different work.
Are there any limitations to your current model that you aim to address in future iterations?
Dealing with a large amount of data with quantum machine learning models is often a bottleneck, which we have found a way around. We will be looking to scale our architecture moving forward to the next seasonal competition phase.
Are there any other AI/ML model components or innovations that you wish to highlight?
Quantum-inspired computing for weather forecasting. World's first deployed and tested algorithm that can be used in real-time for improved forecasting of weather, climate, and extreme events.
Who contributed to the development of this model? Please list all individuals who contributed to this model, along with their specific roles (e.g., data preparation, model architecture, model validation, etc) to acknowledge individual contributions.
Osama Ahmed - Model architecture, scaling, validation, and comparison Sallar Ali Qazi - Data preparation, pre-processing, and model comparison Luca Magri - Model conceptualization, Advising We would like to acknowledge contributions from Felix Tennie for being part of the initial work done and published, for the model development.

Model name

QRNN
Number of individuals supporting model development:
1-5
Maximum number of Central Processing Units (CPUs) supporting model development or forecast production:
< 8
Maximum number of Graphics Processing Units (GPUs) supporting model development or forecast production:
< 4
How would you best classify the IT system used for model development or forecast production:
Single node system

Model name

QROM
Number of individuals supporting model development:
1-5
Maximum number of Central Processing Units (CPUs) supporting model development or forecast production:
< 8
Maximum number of Graphics Processing Units (GPUs) supporting model development or forecast production:
< 4
How would you best classify the IT system used for model development or forecast production:
Single node system

Submitted forecast data in previous period(s)

Please note: Submitted forecast data is only publicly available once the evaluation of a full competitive period has been completed. See the competition's full detailed schedule with submitted data publication dates for each period here.

Access forecasts data

Participation

Competition Period

For the selected competition period, the table below shows the variables submitted each week by the respective team.

Week First forecast window: Days 19 to 25 Second forecast window: Days 26 to 32
Near-surface (2m) temperature (tas) Mean sea level pressure (mslp) Precipitation (pr) Near-surface (2m) temperature (tas) Mean sea level pressure (mslp) Precipitation (pr)

This team did not submit any entries to the competion