University of Southampton student Curtis Parfitt-Ford has been honoured at the TARGETjobs Undergraduate of the Year Awards for action that forced a Government U-turn on the A-level grading algorithm.

The first year BSc Computer Science student spearheaded a legal challenge against August's proposed algorithm that, for class sizes larger than 15, determined grades based on schools' past performance rather than individual achievement.

His powerful efforts, partnered with legal specialists at Foxglove, have been recognised with The Young Activist of the Year Award.

"The A-level algorithm scandal brought issues of algorithmic justice into the public consciousness, with it touching so many lives across the country," Curtis says. "It's an honour and a privilege to have our challenge recognised, and I can only hope that it leads to a greater public conversation and dialogue about oversight of algorithmic decision-making more broadly."

Last summer, UK students did not sit exams because schools were closed during the coronavirus lockdown. In England, grades were decided by the official exam regulator, Ofqual, who proposed an algorithmic model that factored in estimated grades, teacher rankings and schools' past performance.

"This policy was discriminatory," Curtis says, "with bright students in worse-off areas being punished; it was illogical, as the Government had already admitted that teachers were the ones who knew their students' performance best; and it was illegal, as it automatically profiled hundreds of thousands of students across the country without any way of opting out or requesting manual review.

"When I learned that this was happening, the Friday before results day, I knew that I couldn't sit around and let it pass: I could see the impact that it would have on so many across the country. Yet, when I started reaching out to lawyers, I'd never have imagined that we'd force the Government to back down and U-turn in just over a week."

In a matter of days, the campaign raised more than £30,000 to cover potential adverse costs, and more than a quarter of a million people signed a petition Curtis started to Ofqual.

"I'm so incredibly grateful to everyone who signed the petition, who contributed to our crowdfunder, and who helped us make the noise we did about our case," he says.

Curtis is the Technical Officer of the University's LGBTQ+ Society, as well as the Events Lead for its St John Ambulance unit. He also runs a company called Loudspeek, which enables progressive political campaigns to put supporters in touch with their representatives.

"I'm fascinated by the intersection of law and technology, and think the two absolutely have to go together," he says. "Where law lags behind tech, we can see huge injustice, and where tech has to wait for the law to catch up, we risk losing out on revolutionary innovation. Computer science, and the whole Electronics and Computer Science (ECS) field more broadly, offers a huge amount of power - and yet, that power necessarily must come with responsibility.

"As easy as it is for a teenager in their bedroom to invent a website that reinvents the way we interact for good, it's equally easy for people who are genuinely well-intentioned to make systems that are rife with bias, that are inaccessible, or that exacerbate existing inequalities in our society. It's so important that all those implications are thought through and challenged when they're not working how they should - that's the space I'd love to end up working in ultimately."

The Young Activist of the Year Award is partnered by Clifford Chance and One Young World.

Researchers at the University of Southampton are running a series of interlinked projects and consultancy to optimise the design of high voltage underwater cables.

The multidisciplinary team, including high voltage engineering experts in the School of Electronics and Computer Science (ECS), are driving findings that will maximise power transfer and save money in the delivery of future energy.

The research could also lead to a much better understanding of how climate change may be affecting ocean bottom temperatures and, in turn, biological and geochemical processes at the seabed – which are critical to the health of the oceans.

Marine High Voltage Cables (MHVCs) are buried under seabeds around the world. MHVCs for a typical 1 gigawatt windfarm cost about £400 million to design and install, with operation and maintenance costs of several millions of pounds every year.

Professor Justin Dix, of Ocean and Earth Science, explains: "MHVCs primarily operate two ways. Either transferring power between one country and another, for example, hydroelectric power from Norway to Denmark or nuclear power from France to England, or to bring power from our proliferation of windfarms to land. As we explore other sources of renewable energy like tidal power and wave power, the power will need to be brought to land in the most efficient manner.

"If you can better understand the environment the cables are in, and therefore effectively model how heat dissipates away from the cables, you can optimise cables and reduce costs significantly."

The research team has established that, in many cases, cable designs can be changed whilst maintaining a healthy margin of safety.

Dr George Callender, Lecturer in ECS, says: "The current approach to cable rating is effectively based on decades of research into land-based cables. How the seabed environment would alter conventional rating approaches was poorly understood."

The insight is helping to inform the design of SOFIA Windfarm, being built in the central North Sea by RWE. On completion this will be RWE's largest windfarm, providing 1.4GW to 1.2 million homes. The team’s expertise is also being employed in the construction of Inch Cape Windfarm, which will be one of Scotland's largest sources of renewable energy producing 1GW for up to a million households.

Professor Paul Lewin, Head of ECS, says: "We're working with a large number of windfarm companies who want to use our approach rather than just the standard land-based method. We're also working with AP Sensing, a fibre optics company, integrating our ideas into burial-depth prediction software, and most recently we have entered discussion with cable manufacturers such as Nexans."

Looking to the future, the research is also taking a different turn – to understand temperatures at the bottom of the ocean.

Professor Dix says: "The cables strongly record the seasonal variation in ocean bottom temperatures. We think we can back-calculate ocean bottom temperatures from the catalogue of data we have from the last few decades. Ocean bottom temperature is one of the least studied ocean parameters on the planet."

The projects came about thanks to an event organised by the Southampton Marine and Maritime Institute (SMMI).

Professor Damon Teagle, Director of the SMMI, said: "Bringing together academics and researchers from across the University and putting them in a room with a wide range of cable stakeholders, including windfarm developers, cable manufacturers and representatives from the National Grid, was an essential and hugely successful early step.

"As has now been clearly established, the combination of expertise from both earth scientists and electrical engineers is crucial: knowledge of cable design needs to be aligned with knowledge of the ocean floor."

Read the full article in the latest edition of Re:action, the University's research and enterprise magazine.

Global experts in artificial intelligence (AI) and autonomous agent systems are unveiling ground-breaking advances in their field at an online international conference hosted by the University of Southampton.

The 20th International Conference on Autonomous Agents and Multiagent Systems (AAMAS) is examining some of the discipline’s hottest topics including explainable AI, multi-agent reinforcement learning, trust, and AI ethics.

AAMAS is the flagship conference for the autonomous agent systems research community and has been organised this year by several members of Southampton’s School of Electronics and Computer Science (ECS).

Dr Enrico Gerding, Director of Southampton’s Centre for Machine Intelligence, says: "ECS, and the Agents, Interaction and Complexity (AIC) Research Group in particular, has a long history of publishing at this conference, and we are one of the largest groups to work in this field.

"It is a great honour that we were able to win the bid to host the conference in the UK. The plan was to for the conference to be held at the ExCeL conference centre in London, but due to the pandemic it is now fully online. We are greatly looking forward to sharing the latest developments in the field at this internationally renowned forum."

An autonomous agent is an AI system that proactively interacts with its environment, while multi-agent systems consist of many interacting agents - and often people - to solve complex problems.

More than 1,000 researchers and practitioners have registered for this spring's AAMAS conference, with over 300 additional registrations for satellite workshops and tutorials.

Dr Gerding is part of the event's Local Organising Team, with Southampton's Dr Sebastian Stein serving as one of two Local Chairs. AIC researcher Dr Bahar Rastegari is acting as the AAMAS Exhibition Chair, with ECS undergraduate student Balint Gusci supporting as Webmaster.

Former ECS student Lewis Hill has generated a virtual networking space for conference attendees. The Gather Town map is designed with highlights of UK symbols including Big Ben, Stonehenge and a British pub.

The conference opened on Monday and commenced with two days of workshops, tutorials, and the doctoral consortium.

The main conference will advance between Wednesday and Friday and includes sessions with keynote speakers Cynthia Breazeal from MIT Media Lab, Noam Brown from Facebook AI Research, Vincent Conitzer from Duke University and University of Oxford, Orna Kupferman from the Hebrew University of Jerusalem and Shimon Whiteson of University of Oxford and Waymo UK.

Artificial Intelligence technologies from the University of Southampton are helping to deliver more efficient shipping by providing insights, derived from millions of simulations, on the most effective strategies to achieve net-zero targets.

The research collaboration with Shell Shipping and Maritime analyses carbon tax scenarios before predicting the impact of green strategies such as adjusting voyage routes, installing energy efficient technologies and applying carbon offset incentives. The model is intended to provide information and support as shipping companies aim for a net-zero emissions industry.

Southampton AI experts including Professor Gopal Ramchurn and Dr Enrico Gerding are building a large scale agent-based simulation system that models and simulates elements of the shipping industry.

The expanding analysis is part of the University and Shell's partnership work in the Centre for Maritime Futures, which aims to transform the energy shipping industry to be safer, cleaner and more efficient through ground-breaking digital and technological advances.

Professor Ramchurn, Co-Director of the Centre for Maritime Futures, says: "Optimising the transition of shipping fleets to cleaner and more efficient technologies against a backdrop of uncertainty requires trialling a range of future scenarios. We are addressing this challenge through a range of AI and machine learning solutions embedded in a high fidelity simulation platform."

Dr Gerding, Director of Southampton's Centre for Machine Intelligence, adds: "Researchers have created a detailed carbon tax report, highlighting different carbon tax regimes at national, regional and international levels, currently enforced or planned. We developed an agent-based simulation to model different carbon tax regimes, shipping and commodity markets, and provide insights in order to help shipping companies better manage their emissions down."

The simulation system is being built in stages, with each stage adding new features to different apps allowing concurrent and independent development while also ensuring gradual enhancement in functionality, parallel deployment of resources based on their speciality.

To analyse the behaviours and better model the overall simulation system, a detailed data analysis was carried out on data provided by Shell for multiple vessels in their fleet.

The partners' next goal is to ensure that all user requirements are captured to further develop the platform to allow users to run 'what-if' scenarios. This will allow the creation of a multi-user testbed that will enable traders to make better trading decisions and shipping operators to predict the performance of future fleet profiles.

Professor Ramchurn says: "Our work will underpin the design of next generation Uber-freight models that will enable the dynamic routing of vessels and enable a smooth transition to net-zero shipping using green hydrogen both as a fuel and cargo."