List view
Milestone: Systems adapted to intake local regional datasets for finetuned precise forecasts. We will design product output formats per UX research recommendations. Suggested deliverables: - Create forecasts for applications requiring sea ice thickness, sea ice motion – for navigation, migration - Set training schedule to combat data drift (Automation) - Create one framework with Aardvark to cover weather, sea ice, and ocean - Final report and next steps
Due by February 29, 2028Milestone: Full infrastructure handling multi data sources and multi models with multi predictions capability (MIMO). Manuscripts drafted. Suggested deliverables: - Raw data observation (need to determine common format, or who would be the sample case study / actual use case) - Technical paper on the system - Collaborators identified on who can provide long term maintenance
Due by February 26, 2027Milestone: Multimodal capabilities ensembled. UX research completed, findings presented, and white paper drafted. Suggested deliverables: - We aim to start developing a benchmark platform for comparing existing models if we have additional REG support. - Finalise partners who can operationalise this work
Due by August 28, 2026Milestone: Manuscripts drafted, infrastructure designed and in development for multimodal data input and multimodal models. Suggested Deliverables: - Deliver a modular multimodal data ingestion framework - Deliver modules for standardising / normalising data sources - Deliver a modular end-to-end sea ice pipeline in place for proof of concept. - Deliver a prototype of an ensemble learning technique for integration to the pipeline - Base model trained with simple data source - Prepare x research manuscripts drafted for interesting findings leading up to the product. - User Research completed, present findings to drive product outputs. - Pathways identified working in tandem with Aardvark Weather.
Due by February 27, 2026Milestone: Demonstrator in place, additional data source / model in place. Suggested Deliverables: - Present model prototypes (2) and early pipeline capabilities (such as data version control and visualisation for data quality checks) to demonstrate early results to a wider audience at a workshop to be held at Turing. This could be in the form of E&S Seminar or at a joint workshop with UCL/BAS. - Perform initial user research to identify meaningful use cases, present findings. - Determine important outputs beyond sea ice concentration (SIC), sea ice thickness (SIT), sea ice extent (SIE) for daily and monthly forecasts. What this looks like: - Have the demonstrator ready in the first half of milestone three
Due by November 28, 2025•0/4 issues closedMilestone: - Initial prototype of models completed with one input data source and one output prediction in place running off the framework of the existing IceNet architecture. - A minimal viable pipeline with configurable paths developed. - In parallel, one model in place for gridded data with capability to process non-gridded data. - Survey completed for multimodal framework. What success looks like: - Inputs are realistic versions of gridded and non-gridded data (potentially synthetic). - Re-using code from IceNet - The pipeline runs without error - Forecasts recreate training data groundtruth (e.g. OSISAF used for training) better than the persistence model. - The pipeline is flexibly switching between models
Overdue by 16 day(s)•Due by August 29, 2025•6/8 issues closed