David Salvagnini, NASA’s Head of Artificial Intelligence, spoke to Innovation Platform editor Georgie Purcell to discuss how AI technology shapes the trajectory of NASA’s work.
For decades, NASA has used artificial intelligence (AI) technology to support and optimize work across agencies, both on Earth and in space. NASA uses AI to plan and schedule missions for planetary rovers, helping to analyze satellite datasets, diagnose and detect anomalies, develop autonomous systems, and more.
Some of the AI tools used by NASA include machine learning, which uses data and algorithms to train computers on creating classifications, formulating predictions, and discovering similarities or trends across large datasets. Using AI tools offers a range of benefits that will turn your game into NASA jobs, including streamlining decision-making, saving resources, and increasing workforce efficiency.
To learn more about NASA’s AI activities and what the agency’s future holds for such technology, Georgie Purcell spoke with David Salvagnini, NASA’s Head of Artificial Intelligence.
What role does AI technology play in NASA’s work?
NASA is very involved in the use of artificial intelligence and machine learning in several ways, especially for more traditional means. Mainly, these technologies are used in our scientific work, helping to discover objects in the solar system or in the distance. One example is the ability called Exominer. It is trained to identify exoplanets, planets and even distant solar systems using machine learning-trained AI. Interestingly, this used data was collected 15 years ago and was able to identify previously undiscovered objects through AI models.
Another area where the use of AI technology is common is autonomy. For example, Mars’ perseverance must safely navigate the surface of Mars and avoid obstacles and dangers, whilst dealing with the potential risk of delays in communication between Mars and Earth. As you might see in self-driving cars in today’s market, onboard AI systems help rovers handle the environment around them and render decisions about how the vehicle operates. If the rover had to communicate with controllers on Earth, the autonomous nature of the system removes the propagation delay problem. This technology has been frequently tested as part of the systems engineering process in this NASA, reducing and reducing the use of AI in this case.
In collaboration with IBM, we recently released the AI Prithvi-Weather-Climate Foundational Model for a variety of climate and climate use cases. This model provides a flexible and scalable way to address the myriad challenges associated with forecasting short-term weather and long-term climate. The resulting data is published from the model, so anyone can access and use it. Many peer reviews and collaborations are involved in this type of work.
The use of AI in space is very exclusive. In traditional AI that businesses may use, large computing backends like cloud providers are the processing engines of much of the work. Spaces do not have access to the cloud platform. When we think about the future of AI and how to continue to enable AI in space-based systems operating in extreme environments, we need to explain several different factors. Such include the extreme environment itself, the radiation exposure to electronic equipment, power budgets and restrictions. Unlike here on Earth, when it comes to space-based vehicles, it’s not as easy as buying more electricity. There are strict budget and weight restrictions to pay respect. There are many unique elements in the way NASA uses AI, especially to support space-based missions, creating particularly complex challenges for us. However, NASA has a long history of overcoming these challenges.
What potential does AI have for the future of NASA’s mission?
One area in which AI plays an important role is making things more adaptive and autonomous. For example, on a typical aircraft today, there is a crew member, but there is also an autopilot system that controls the plane for most of the flight. This technique is highly deterministic – if the speed changes, a change in control is made. Similarly, NASA is currently working to enable dynamic adaptation to a variety of environments, whether on orbit or on the surface. We are working to gain much higher situational awareness about the environment, which could be fed into autonomous systems. Autonomous systems can more reliably execute responses to a more dynamic set of conditions.
Orbital debris is a major problem for NASA, and concerns are growing as space activity continues to rise rapidly. We recently released a Space Sustainability Strategy to measure and assess the sustainability of Earth, Earth Orbit, Cislunar Space and Deep Space. AI plays an important role in this not only detecting orbital debris, but also in some of the actions the system may take to repair it.
Additionally, AI is used in adaptive communication systems. Communication is considered to be something like a telephone as a single mode. However, space environments show more complex needs when making calls using multiple communication networks. AI can be used to understand the conditions at some point and select the optimal communication path and timeliness requirements for data being sent based on the conditions. AI plays a major role in optimizing communication and improving communication reliability.
It is also seen that AI models interact with other AI models. Individual AI models with individual understanding of different parameters can cooperate to raise awareness of factors such as climate, for example. There is a lot of excitement about that possibility.
There are also many personal and daily activities that NASA pursues in the field of AI. For example, NASA is looking at ways that AI can support the workforce, from image development to summary notes from meetings. It is important to see AI as a tool that makes our work easier and more efficient, rather than as a threat to replace human work.
Can you provide some recent examples of projects focused on AI technology?
The Prithvi climate model mentioned earlier is a great example.
Another example is to use AI to design structural components. AI allows for much more liquid and nonlinear designs than humans have made. AI can also generate components much faster than humans, but when components are stress-tested, they can achieve equal or better results than human-generated versions.
The Mars Perseverance Rover and ingenious Mars helicopter that flew to the surface of Mars are important examples of how AI adapted to help us in our research. Furthermore, we are also considering long-term human spaceflight beyond the moon. For example, we are developing technologies designed to deal with medical crew emergencies. Typically, astronauts are not doctors and may not be able to contact Earth’s experts to help out in cases of factors such as a power outage. It is important to equip them to deal with emergencies in real time. We are working to provide our crew with AI capabilities to assist in the rapid diagnosis of medical conditions and to suggest potential therapeutic effects.
What challenges did you encounter when implementing AI technology? How are these dealt with?
As mentioned earlier, space-based systems that operate in extreme environments and are exposed to radiation, for example, have electronic equipment issues. There are also restrictions on power budgets and weight. They are a major challenge that we don’t need to think about as AI on the planet. There is also the additional challenge of having to bring that computing into it. It’s not just reaching out to the cloud to process large amounts of data. Spaces pose a completely different and unique set of challenges.
There is a lot of fear surrounding AI implementation. Naturally, people are concerned about ethical and responsible use, privacy, transparency, and more. It takes time to get used to the technology and the possibilities it offers. Thankfully, unlike other federal agencies, NASA is not involved in providing public services, such as processing health benefits claims. Such agencies have more work in addressing responsible, privacy and ethical use concerns that exist.
There are cultural difficulties too. Like other organizations, we need to address workforce upskills and reskills. Some people are eager to jump into AI at full speed, while others are very reserved about it. Every organization has skeptics who are a little worried about new things and are not yet comfortable. If I don’t describe it as a challenge, I will be tolerant. We must meet people where they are, be equipped to use those tools effectively, know which tools are needed, and help them use them responsibly. It will take time for the entire workforce to be fully prepared to embrace everything AI has to offer. But I have to emphasize that NASA’s culture is very pioneering. When people at NASA see opportunities that can be effective with the use of technology, they often want to embrace it. Adopting new technologies involves risks, but NASA is extremely effective in managing risk. NASA balances innovation and wise risk management, especially when adopting and implementing AI, especially when complementing the entire workforce with generated AI tools.
This article will also be featured in the 20th edition of Quarterly Publishing.
Source link
