HPE here to help on the occasion of World Earth Day
The 04/2020 – Hewlett Packard Entreprise
HPE reiterates its commitment to the environment on the occasion of World Earth Day
22 April 2020, at a time when climate change represents one of the greatest challenges for the future of humanity, HPE brings its expertise in high-performance computing to support researchers in the development of advanced artificial intelligence methods for observing the environment.
HPE’s strategy for sustainable development
Our climate strategy affects all areas of our organisation and is essential to our business. As a technology company, HPE has a key role to play in responding to climate change – both through transformative efficiencies and by building adaptive capacity.
“We have set strategic goals to minimise our environmental footprint across our entire value chain, focusing on areas where our impact is greatest,” comments Mateo Dugand, Head of Sustainability EMEA at HPE. “I am proud of the work our teams are doing with the scientific community to help us preserving our environment”.
HPE was the first IT company to set science-based targets (SBT) to reduce GHG emissions across its entire value chain (operations and supply chain)
These targets are endorsed by the Science Based Target initiative and are in line with the recommendations of the internationally recognised Paris Climate Accord, which aims to limit the rise in global average temperature to 1.5°C in order to significantly reduce the risks and effects of climate change. In 2019, HPE has set a new target to reduce our operational GHG emissions by 55% from 2016 levels by 2025.
Studying the impact of climate change on fauna and flora in the Alps
MIAI, the Grenoble-based interdisciplinary institute for Artificial Intelligence created in September 2019, develops the new generation of AI models and systems, from hardware and embedded architectures to software, with a particular focus on the environment. HPE brings its expertise in high-performance computing and assists researchers in porting their code to high-performance machines enabling them to move on to large-scale analyses.
Within this institute, Jocelyn Chanussot, Professor at the GIPSA-Lab laboratory in Grenoble-INP and Wilfried Thuiller, Director of Research at the Alpine Ecology Laboratory (LECA) are leading a chair dedicated to the development of advanced artificial intelligence methods for the observation of the environment and more specifically alpine ecosystems, which are particularly affected by global warming. In order to study this phenomenon, 40m2 of grasslands were transplanted in 2016 and have since been monitored by in situ observations (morphological characteristics, photosynthetic productivities, changes in microbial composition by environmental DNA) and since 2020 by traditional (RGB) and hyperspectral photographs at very high spatial (~1cm) and temporal (~1 time per week) resolution.
In order to measure, quantify and understand the impact of climate change on our environment, many means of observation are available: satellites, airborne campaigns, the use of drones, in-situ measurements, and even laboratory measurements for DNA analysis of plants and animals. All these observations generate a very large volume of data (>100TB), the so-called “Big Data”, with very different scales of observation. “The success of these projects is due to the synergy between ecologists, data scientists and high-performance computing experts. The MIAI Institute has played a key role as a catalyst.” explains Jocelyn Chanussot.
“We are very proud to be involved in this research project at the 3IA in Grenoble (MIAI), not only for its interest in the fight against global warming, but also to refine our skills in image analysis and the resulting learning models.” commented Pierre Hoffer, Director Presales EMEA at HPE for HPC and AI solutions.
Supercomputers such as GENCI’s Jean Zay machine, designed by HPE and inaugurated in January 2020, are in fact indispensable for training processing algorithms on colossal volumes of data and thus extracting information from this highly heterogeneous mass of data.
Collecting sound data at the edge of palaeotropical rainforests and arid deserts
A team of ecologists, led by Dr. Brian Pijanowksi of Purdue University in the United States, is using data from the earth to study indicators of global change. Dr. Pijanowski and the team from Purdue University’s Center for Global Soundscapes, also known as “Soundscape”, ventured into the wild to collect samples of earth sounds that, through collaboration with HPE, would provide valuable information that will lead to solutions to preserve the planet’s biodiversity for future generations.
Sound samples related to global warming can be heard in birds singing at higher tones in more urbanized environments and in the more frequent occurrences of thunder in Fairbanks, Alaska, which are now triggered by changes in climate patterns.
The Soundscape team began their global listening tour in a tropical forest in Borneo, one of Asia’s largest islands, where they placed battery-powered acoustic sensors and recorders to capture the sound. The process continued in a number of remote areas, collecting a significant amount of audio data from 80% of the world’s major ecosystems. This represents a ton of different sounds from all kinds of activities such as animals, weather and moving water.
The Soundscape team needed a combination of powerful computing, deep learning, and data mining and analysis solutions to ingest the data and process it to learn from it. The team was able to extract large amounts of complex data from remote sensors and recorders that were strategically dispersed across the wilderness, using high-performance, state-of-the-art computing solutions from Edgeline Converged HPE systems and ProLiant HPE servers. Once the fragmented data was consolidated, it was processed and analyzed for data visualization using Apollo HPE systems, which are ideally suited to support in-depth learning and analysis capabilities.