Chatbots are saving the auto insurance industry during the coronavirus pandemic, powered by machine learning...
Chatbots are saving the auto insurance industry during the coronavirus pandemic. Powered by machine learning (ML), digital insurance platforms research applicants’ driving records, analyze data, apply risk metrics to coverage and pricing, and issue policies without face-to-face interaction.
Covid-19 has accelerated the development of machine learning applications that solve today’s problems, according to Professor Daniela Rus, faculty director for MIT’S Computer Science Artificial Intelligence Lab (CSAIL) research alliance. Last week, CSAIL launched, via webinar, MachineLearningApplications@CSAIL to develop applications for the latest ML technologies, research challenges limiting ML, and provide professional development for the digital workforce.
Businesses that support CSAIL research stand to immediately benefit from its application. Chatbots already use ML to recommend products, optimize pricing and create classifications within systems. More broadly, enterprises use ML for energy savings and predictive analytics. Restrictions on face-to-face commerce due to Covid-19 is a significant concern for retailers and businesses. Machine learning and AI have the potential to transform business processes, Rus said.
CSAIL’s “no idea is too crazy” approach has already yielded numerous breakthroughs in computing technology. MIT’s AI efforts date back to 1959 and its AI Lab pioneered methods in image-guided surgery, language-based web access, micro displays and robotics. CSAIL is the merger between the AI Lab and MIT’s Laboratory for Computer Science (LCS), founded in 1963. The LCS is known for the development of the Compatible Time-Sharing System (CTSS) and Multics.
CSAIL currently oversees more than 60 research groups working on hundreds of projects.
Machine learning can be applied across business segments from grocery stores to 5G networks by harnessing data automatically generated through commerce. Machine learning uses data to derive rules or to predict future action. The ability to digitize transactions is paramount to a healthy economy, said experts on CSAIL’s ML introductory webinar.
Data in isolation is generally not helpful to businesses, Rus explained. “Data is only as good as its usage — so you need to test and verify data and think hard about what process the testing system aligns with. Think about use cases and what you can derive from simulations and how you can apply those in real life. ML will make mistakes so you need the machines and people to work together. Ultimately, humans are responsible for decisions.”
“If you ask what keeps executives up at night, the answer will be using ML or AI to grow their business no matter what industry they’re in,” Rus added. “The opportunities for using AI and ML are huge, and the time to pursue them is now.”
Commercializing technology as quickly as possible is the goal of CSAIL and the businesses that support its research. Arrow Electronics Inc., Cisco, Retail Business Services (Ahold Delhaize) and SAP are founding members of the ML initiative and envision applications in quality control, supply chain management, connectivity and forecasting, among others. (Full disclosure: Arrow is the parent company of EE Times’ publisher, AspenCore Media.)
The businesses that support ML research, Rus said, can:
EE Times asked Arrow how much it invested in becoming one of the founding members of MIT’s CSAIL research alliance. The company did not respond by press time.
Global spending on artificial intelligence (AI) will reach $97.9 billion in 2023, according to research firm IDC, more than two and one half times the $37.5 billion spent in 2019. The compound annual growth rate (CAGR) for the 2018-2023 forecast period will be 28.4%.
Healthcare has become a priority during the pandemic, and machine learning is condensing medical research from years into months, said Rus. “The fact that we are talking about a [Covid-19] vaccine in terms of months, rather than years, is due to the acceleration that can take place thanks to machine learning,” she said.
Collaboration between humans and AI increases diagnostic accuracy, according to Rus. One radiology study asked doctors and machines to identify, from scans, cancerous from non-cancerous tissue. The human error rate was 3.5%; machines erred 7.5% of the time, but together the error rate was reduced to 0.5%.
“Imagine a point in time where doctors in rural settings, or those that are overworked, won’t have to be on top of the latest recommendations and studies. They can provide data on the patient in front of them and ML or AI can customize medicine for fewer side effects on a body. Imagine a world where ML and artificial intelligence (AI) will help with cognitive and physical tasks much like smartphones provide help with computing and number-crunching,” Rus said.
Roadblocks to widespread ML adoption include human to machine interface, Rus said. “Chatbots simplify processes but the interfaces are still clunky for people that don’t live on their PCs.” Language translation remains problematic. ML matches patterns to translate from one language to another, but doesn’t measure context or intent, she said. At the same time, chatbots retain data and can answer questions more quickly than humans.
“Machine learning is all about past data; AI is about decision-making, and you can include robotics, which is computation in motion. These can make a huge difference in how business and individuals solve problems,” Rus concluded.