How to scale the power of AI simulation
source link: https://toptech.news/how-to-scale-the-power-of-ai-simulation/
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
How to scale the power of AI simulation
We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 – 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!
AI-driven simulations are suddenly everywhere. From digital twins that provide insights into current performance to advanced simulation intelligence that evaluates future scenarios, organizations in nearly every industry are taking advantage of the evolution of AI simulation to speed up operations, reduce risk and boost better decision-making.
And almost all companies are on board the AI simulation train. According to PWC’s 2022 AI Business Survey, it is almost unanimous – a whopping 96% of survey respondents said they plan to use AI simulations this year. Meanwhile, 57% of those deemed AI leaders plan to use simulations to forecast market conditions, while 54% of AI leaders plan to use simulations to support financial, sales and marketing planning as well as enhance supply chain and operations.
This fervent focus on the business power of AI simulations is not surprising, Anand Rao, principal with PwC’s U.S. advisory practice, told VentureBeat.
“A few years ago, some of this was not possible because there was not enough compute power or enough data,” he said. “But now you can simulate millions of customers, for example, and the details around how they are making decisions – and then you can get tangible output to tailor and target your products and services.”
Still, Rao said while he wasn’t surprised that the number of respondents using or planning to use AI simulations had increased, he was surprised by how much. “The pandemic changed things dramatically,” he said. “Digital twin modeling or other AI simulations work particularly well when you don’t have a lot of historical data and when you have a very uncertain future environment – which is what happened during the pandemic.”
How to sustain AI simulations at scale
The question is, how can companies go beyond a one-and-done AI simulation in a crisis to sustainable efforts at scale across the organization? Rao suggests three key actions:
- Bring your business and technical people together. “From a business perspective, there are very few people with the right talent who can do this,” Rao said. Simulation specialists, who typically have an engineering background, need to be brought together with data scientists and business leaders. For the right simulation, “you need to understand the business as well as how to build these models with specific tools and techniques,” he added.
- Make simulation part of the overall technology fabric. By integrating AI with the cloud and making digital twins a platform capability, organizations can use AI’s power effectively to create business-relevant simulations, Rao said. “The technology that has evolved has been primarily focused around machine learning and deep learning, not so much on simulation,” he said. “Simulation tended to be more of a niche area that some engineers were using, but it was never part of the overall technology fabric. Now companies are saying, ‘Hey, I want the simulation capability within my IT stack, I want to be able to pull the data from all of my existing data stores, I want to connect it back to my Salesforce system and operational system so that this becomes part of the way I make decisions.’”
- Create synthetic data. Simulation models can create data to train machine learning models, which require huge amounts. Simulations can generate images of faces with multiple angles, contrast levels and brightness for facial recognition models, for example. In addition, synthetic data can be refined. “You can start with existing data you have about your customers and enrich it with external or synthetic data,” said Rao. “As you see what is happening in the real world, you can revise it, so the data becomes much richer.”
AI simulations will continue to trend upward
According to Rao, enthusiasm for AI simulations will continue to trend through 2022 and beyond.
“During the pandemic, people were looking for something that would keep them from being in the dark about everything from the supply side to the demand side to personnel,” he said. “When there is that kind of uncertainty in the market, these techniques are very useful.”
VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn more about membership.
Author: Sharon Goldman
Aggregate valuable and interesting links.
Joyk means Joy of geeK