two-days of learning, fun and networking - with artificial intelligence and SUSTAINABILITY enthusiasts from across the world

To join our social as an attendee, you would need to be a registered participant at ICLR 2022 - see details here. Join us here - The entire event will be digital and take place on Zoom.

Background of the event

With rapidly rising carbon emissions globally, it is the need of the hour to transition to clean energy sources – such as wind, solar etc. However, renewable energy sources like wind turbines are complex engineering systems that regularly suffer from operational inconsistencies and failures, leading to downtimes and energy production short of the full potential. AI can help support the operations & maintenance (O&M) of such energy systems, helping predict incipient failures and also suggesting maintenance actions to fix/avert faults. This can help bring down O&M costs as well as reduce downtimes and increase availability of the energy systems. At present, there is very limited focus on leveraging AI in the renewables domain.

This social will aim to emphasise the opportunities (e.g. fault prediction, suggesting O&M activities, power forecasting etc.) and challenges (e.g. data confidentiality in the industry, lack of historical failure data etc.) in applying AI for smoother transition to clean energy, with various avenues such as transfer learning and natural language generation to ensure trustworthy and reliable decision support. This social will focus on a short presentation from the organisers, followed by a panel discussion from invited experts on "AI for renewable energy" followed by open Q&A on the first day (25th April, 11 am - 1 pm GMT). We would have a fully informal/friendly socialising session on challenges and opportunities in energy transition, along with an invited talk on the role of AI in climate change mitigation - on the second day (27th April, 11 am - 1 pm GMT). See full schedule of the event here.

key themes OF the social

• The need of the hour in transitioning to low-carbon energy sources for realising the United Nations' SDGs - particularly SDG 7 (Affordable and Clean Energy).

• The practical challenges faced by the renewable energy industry - high costs in operations & maintenance (O&M), unexpected failures and downtimes, varying availability and intermittent power production (e.g. during snow, low-wind speed etc.), energy storage costs, challenges in integration with the electric grid and so on.

• How can AI help - from a sustainability perspective as well as an industrial perspective? The potential for digitalization to be a disruptive game-changer in the renewable energy domain - thus providing affordable and reliable energy to households and industries in global cities and communities.

• Why do we not see much AI in the renewables domain? Challenges including data availability and quality, lack of transparency in black-box natured AI models (especially deep learners) and issues in deploying models for real-time decision support in the industry etc.

• Potential for leveraging Explainable AI techniques and knowledge transfer mechanisms - what role can natural language generation, transfer learning, deep reinforcement learning techniques etc. play beyond current practices in using predictive AI models e.g. for power forecasting and fault prediction?

• Ethical and moral concerns in applying AI in the renewables domain - widening of digital divide between the developing and developed world, environmental effect of training large-scale AI models etc.

• What would be a viable roadmap to see more AI in the future in the renewables domain - role of industry, academia and governments to work hand-in-hand, large-scale sharing of open data, focus on Explainable AI, greater knowledge exchange between the developing and developed world etc.