Artificial Intelligence (AI) Lab

The British Antarctic Survey’s Artificial Intelligence (AI) Lab is at the forefront of innovative research and innovation, employing AI and machine learning techniques to tackle a broad spectrum of challenges within the institute’s remit. From sea ice forecasting and space weather, ice dynamics and sea level rise, seasonal polar operations and automation, to tracking wildlife and icebergs from space and benthic biology on the seafloor.

Extracting knowledge from the diverse range of datasets and domain areas which span various spatial and time scales poses significant challenges. The BAS AI Lab leverages the power of AI to intelligently integrate different types of data to pave the way for new discoveries and unlocking previously inaccessible information.

One of the AI Lab’s focuses is optimising data collection processes in remote and hostile environments to reduce carbon emissions and maximise science output. By integrating satellite observations, surface sensors, and optimising task planning for infrastructure and logistics (including ships and autonomous marine vehicles) the AI Lab strives to streamline data acquisition and improve the efficiency of scientific expeditions. This approach ensures that valuable resources are used sustainably, minimising carbon footprint.

BAS AI Lab Themes
Research areas within the BAS AI Lab

As a leading national entity, BAS’s AI Lab actively collaborates with national and international partners to co-develop and co-deliver AI methods that are tailored to meet the unique requirements of polar research. By working closely with experts from diverse backgrounds, the lab ensures that the developed AI techniques are fit for purpose and aligned with global standards to enable wider adoption.

An important aspect of the AI Lab’s mission is to put novel, freely available digital solutions into the hands of decision-makers. By leveraging the power of AI, the lab creates operational tools and technologies that aid policymakers and stakeholders in making informed decisions to safeguard our planet. These solutions provide valuable insights and enable proactive mitigation and adaptation measures to help address the climate and biodiversity crisis.

PhD opportunities

 

habbot

Harrison Abbot

Research Scientist

user_3529-2

Ellen Bowler

Machine Learning Research Scientist

jambyr

James Byrne

Research Software Engineer Lead

gecoomb

George Coombs

Research Scientist

user_3472-2

Ben Evans

Machine Learning Research Scientist

aykry

Ayat Fekry

Research Scientist

marfox

Maria Fox

Principal Researcher in Environmental AI

user_2834-2

Rachel Furner

PhD Student

samhall

Samuel Hall

Research Scientist

jask

Scott Hosking

Environmental Data Scientist

anddon76

Andrew McDonald

PhD Student

user_3491-2

Martin Rogers

Machine Learning Research Scientist

jonsmi

Jonathan Smith

Principal Research Scientist

kenzi22

Kenza Tazi

PhD Student

mior

Michael Thorne

Research Scientist

donyaz

Donya Yazdani

Research Scientist

Digital Twins of the Polar Regions

Digital Twinning is next generation technology for data fusion and computer modelling enabling us to rapidly get answers to “what-if” questions. DTs are already in operation in industry and involve …

IceNet

IceNet is a probabilistic, deep learning sea ice forecasting system developed by an international team and led by British Antarctic Survey and The Alan Turing Institute [Andersson et al., 2021]. …


AI for smart conservation

In the AI for smart conservation project, BAS are collaborating with local ecologists and conservation agencies to develop decision-making tools informed by sea ice forecasts. By combining satellite observations, GPS …

DI4EDS

Environmental research relies on digital infrastructure (hardware, software and methods) to provide services that help researchers answer questions about the environment around us, and innovators to work out ways that …

Using AI to track icebergs

23 November, 2023

Researchers are using a new AI tool to detect icebergs in the Southern Ocean. This is the first step towards scientists being able to track the complete life cycle of …


AI tool to revolutionise polar ship navigation

15 November, 2022

Artificial Intelligence (AI) will enable ships navigating in polar ocean conditions to be more efficient using a new route planning tool created by British Antarctic Survey (BAS) researchers. The tool …




Using AI to track whales from space

4 February, 2021

British Antarctic Survey (BAS) scientists will work with an Artificial Intelligence company after being awarded a contract from the Canadian Space Agency (CSA) to support the protection of an endangered …


PhD centre will nurture new leaders in Earth observation

9 January, 2020

A new centre will enable 50 fully-funded PhD researchers to harness satellite data to tackle global environmental challenges. The Centre for Satellite Data in Environmental Science (SENSE) will bring together expertise in …


Using AI to help tackle global environmental challenges

26 February, 2019

A new Centre for Doctoral Training, involving researchers from British Antarctic Survey, will develop Artificial Intelligence (AI) techniques to address critical environmental challenges. Climate change and environmental hazards pose some …


Watching whales from space

1 November, 2018

Scientists have used detailed high-resolution satellite images provided by Maxar Technologies’ DigitalGlobe, to detect, count and describe four different species of whales. Reported this week in the journal Marine Mammal …











Review of Satellite Remote Sensing and Unoccupied Aircraft Systems for Counting Wildlife on Land

8 February, 2024 by Ellen Bowler, Hannah Cubaynes, Marie Attard, Penny Clarke, Peter Fretwell, Richard Phillips

Although many medium-to-large terrestrial vertebrates are still counted by ground or aerial surveys, remote-sensing technologies and image analysis have developed rapidly in recent decades, offering improved accuracy and repeatability, lower…

Read more on Review of Satellite Remote Sensing and Unoccupied Aircraft Systems for Counting Wildlife on Land

Unsupervised machine learning detection of iceberg populations within sea ice from dual-polarisation SAR imagery

1 November, 2023 by Andrew Fleming, Anita Faul, Ben Evans, David Vaughan, Scott Hosking

Accurate quantification of iceberg populations is essential to inform estimates of Southern Ocean freshwater and heat balances as well as shipping hazards. The automated operational monitoring of icebergs remains challenging,…

Read more on Unsupervised machine learning detection of iceberg populations within sea ice from dual-polarisation SAR imagery

Histopathological screening of Pontogammarus robustoides (Amphipoda), an invader on route to the United Kingdom

1 September, 2023 by Martin Rogers

Biological invasions may act as conduits for pathogen introduction. To determine which invasive non-native species pose the biggest threat, we must first determine the symbionts (pathogens, parasites, commensals, mutualists) they…

Read more on Histopathological screening of Pontogammarus robustoides (Amphipoda), an invader on route to the United Kingdom

Using Probabilistic Machine Learning to Better Model Temporal Patterns in Parameterizations: a case study with the Lorenz 96 model

10 August, 2023 by Scott Hosking, Raghul Parthipan

The modelling of small-scale processes is a major source of error in climate models, hindering the accuracy of low-cost models which must approximate such processes through parameterization. Red noise is…

Read more on Using Probabilistic Machine Learning to Better Model Temporal Patterns in Parameterizations: a case study with the Lorenz 96 model

Self-Recover: Forecasting Block Maxima in Time Series from Predictors with Disparate Temporal Coverage Using Self-Supervised Learning

1 August, 2023 by Andrew McDonald

Forecasting the block maxima of a future time window is a challenging task due to the difficulty in inferring the tail distribution of a target variable. As the historical observations…

Read more on Self-Recover: Forecasting Block Maxima in Time Series from Predictors with Disparate Temporal Coverage Using Self-Supervised Learning