Reflecting on the significance of Earth Day in 2022, it is becoming very clear that only an approach rooted in deep science will enable us to tackle the long list of challenges posed by the changing climate.
Of course, there’s a lot in the most recent Intergovernmental Panel on Climate Change (IPCC) report to inspire feelings of despondency – that humanity has backed itself into a global climate crisis corner. Yet, as IPCC Chair Hoesung Lee states in the opening of the report: “We have the tools and know-how required to limit warming.”
To this end, IBM Research and our partners worldwide are enhancing and applying the latest in Artificial Intelligence (AI), hybrid cloud and other emerging technologies to The world is changing rapidly every day, and the way we used to solve problems won’t cut it anymore. At IBM Research, we’re combining our expertise in quantum computing, AI, and hybrid cloud to drastically increase how quickly we can discover solutions to tackle today’s most urgent problems. Read more about the Accelerated Discovery team's work.accelerate the discovery of solutions that can mitigate climate change and help the world adapt to increasingly severe climate conditions. Our goals are to create more energy-efficient computer hardware and software, make infrastructure more resilient to extreme climate events and protect the riches of the planet, both biological and man-made, with the help of geospatial analytics.
The development of more powerful, yet highly energy efficient, hardware and software serves two important purposes. First, these systems are the engines to help accelerate the scientific discovery required to develop timely climate solutions such as discovering new materials that can be used for carbon sequestration. Second, more energy-efficient IT systems help keep data centers themselves from contributing to the problem.
We have the tools and know-how required to limit warming.
The total growth in energy consumption for computing out-paces the growth in global energy production.1 Factors that are contributing to this growth rate include an explosion in the amount of information data centers must process, store and transfer, as well as the emergence of energy intensive AI workloads.
“We must take a holistic and systematic approach to reduce the carbon footprint associated with computing systems” said IBM Fellow Tamar Eilam.
“In today’s data centers, we can reduce the overhead of cooling and power distribution; we can maximize the use of renewable energy for computing by coupling workload and grid dynamicity; and we must work on new innovative digital accelerators for specialized workloads, and even entirely new computational models that drastically reduces energy such as Analog AI. In addition, we ought to also address the embodied server emissions through innovations in materials and processes.”
To reduce AI hardware’s impact on the environment IBM Research is developing AI chips that rely on reduced precision to cut the energy consumption while maintaining model accuracy. This includes AI accelerator chips built with 7nm technology2 that we are making available at our The IBM Research AI Hardware Center is a global research hub headquartered in Albany, New York. The center is focused on enabling next-generation chips and systems that support the tremendous processing power and unprecedented speed that AI requires to realize its full potential. Learn more.AI Hardware Center in Albany, NY. These chips incorporate ultra-low precision hybrid 8-bit floating point number (FP8) formats for deep-learning models.
Recently, IBM Research also demonstrated an approach for energy efficiency3 using low voltage DNN accelerators. While low voltage is desired for energy efficiency, an undesired outcome is that it can cause bit-flipping in the SRAM. Our team demonstrated bit-error robustness in low voltage chips through a combination of robust fixed-point quantization, weight clipping, and random bit error training (RandBET).
Another aspect in battling climate change is helping with the energy transition to an enhanced, resilient, and climate-aware electric grid powered by renewable energy sources. For electric utilities companies that manage such critical infrastructure, this often means accounting for the extreme weather events that are becoming more frequent and more severe, vegetation encroachment on remote equipment, and other disruptions.
IBM has partnered with Irth Solutions to provide cloud-based software-as-a-service (SaaS) solutions that automates processes for damage prevention and asset protection across critical network infrastructure assets. Irth taps into IBM’s geospatial, weather, socioeconomic, news, and other insights to enhance its platform. This combination creates a robust 360-degree situational awareness of what might impact an asset so Irth’s customers can determine the best action to protect or minimize damage to the delivery of essential services.
“We’re partnering with IBM to provide our customers with better abilities to protect their critical network infrastructure,” said Brad Gammons, CEO, Irth Solutions. “Being able to understand the condition and integrity of assets protects the environment by keeping an oil leak from occurring or detecting a methane leak earlier.
“Irth protects Earth today and the infrastructure that delivers decarbonization through electrification in the future,” Gammons adds. “The electric grid must be unbelievably resilient for the energy transition to a decarbonized world. As electrification expands, the network becomes vastly more complex to operate and requires that it is connected and monitored 24/7 in real time. Therefore, the telecommunications network must be more reliable than ever.”
Rising global temperatures and unstable weather created by climate change threaten not only modern infrastructure but also heritage sites dating back hundreds or thousands of years.
IBM Research has worked with Japan’s Yamagata University for the past several years to help researchers there find and study new geoglyphs that form the mysterious Nasca Lines in southern Peru. Since 2006, Yamagata researchers have discovered more than 100 new geoglyphs in an area covering about 500 square kilometers. More recently, the researchers incorporated the IBM PAIRS platform and AI to not only search for new geoglyphs but also to identify climate threats to these ancient geoglyphs, which date back about 2,000 years.
“Some of the Nasca geoglyphs, a UNESCO World Heritage site, are in danger of destruction due to water runoff caused by recent heavy rains,” said Yamagata Professor Masato Sakai. “In order to protect these geoglyphs, it is necessary to analyze vast amounts of topographical, meteorological, and archaeological data, which is why working with IBM Research by utilizing IBM PAIRS is so important.”
IBM is proud to prioritize an agenda that delivers to the world the most cutting-edge technologies that also advance the health and safety of our planet. And, we believe it’s on everyone — from researchers and technology leaders, to government and society at large — to push forward new and lasting discoveries that can combat climate change and elevate sustainable solutions. On this Earth Day, we share the message of hope, that the scientific method and rigorous research can help address negative impacts changing climate is having on our communities and businesses.
Learn more about what our team is doing in the space of climate and sustainability. And learn more about IBM Impact, a new framework for our company’s environmental, social and governance (ESG) work, including our 2021 ESG report.
- Semiconductor Research Corporation. The Decadal Plan for Semiconductors: A pivotal roadmap outlining research priorities. https://www.src.org/about/decadal-plan/.↩
- A. Agrawal et al., "9.1 A 7nm 4-Core AI Chip with 25.6TFLOPS Hybrid FP8 Training, 102.4TOPS INT4 Inference and Workload-Aware Throttling," 2021 IEEE International Solid- State Circuits Conference (ISSCC), 2021, pp. 144-146, doi: 10.1109/ISSCC42613.2021.9365791.↩
- Stutz, D., Chandramoorthy, N., Hein, M., Schiele, B. Bit Error Robustness for Energy-Efficient DNN Accelerators. Part of Proceedings of Machine Learning and Systems 3 (MLSys 2021).↩