Issue link: https://resources.mouser.com/i/1437726
15 INTEL 2020 are both predictive and prescriptive, meaning they tell you what's likely to happen and recommend steps you can take to address emerging issues and influence outcomes. Let's look at some specific use cases for artificial- intelligence-driven predictive applications across a range of industries. Enabling predictive maintenance In many industries, predictive systems driven by machine-learning techniques are helping operators keep equipment up and running at an optimal performance level while reducing maintenance costs. These systems monitor the performance and condition of equipment to anticipate failures and enable proactive maintenance. A few examples: • Smart manufacturers are using AI systems in conjunction with data from sensors and the Internet of Things to predict - machine failures. The goal is to use predictive maintenance to avoid issues on the manufacturing line, resolve problems quickly and proactively, and minimize disruption to operations. • Wind-energy producers are using AI systems in conjunction with data from sensors and the IoT to predict the likelihood of wind-turbine failures and proactively address issues that might arise. • Telcom providers are using machine- and deep- learning systems to guide preventative and predictive maintenance-related actions to reduce downtime of mission-critical systems, such as telephone billing clusters. The payback for these applications can be huge. A report by research firm McKinsey & Company notes that AI-driven predictive maintenance can increase asset productivity by up to 20 percent and reduce maintenance costs by up to 10 percent, while greatly reducing machine downtime caused by maintenance work. Predicting healthcare outcomes In healthcare settings, the ability to predict the likelihood of patients developing certain complications and conditions can help clinicians work proactively to prevent problems and improve patient outcomes. Penn Medicine, which operates a network of healthcare facilities in Pennsylvania and New Jersey, proved this point by using a collaborative data science platform it created with Intel. In its first trials of the platform, the healthcare provider developed algorithms to help predict and prevent two of the most common and costly issues for hospitals: sepsis and heart failure. The results were amazing. For example, Penn Medicine was able to correctly identity about 85 percent of sepsis cases and to make these identifications as much as 30 hours before the onset of septic shock. These AI-driven results were far better than the expected outcomes with conventional methods. With these more accurate and timely predictions of the sepsis risk, clinicians can deliver treatments sooner, speeding time to recovery for the patient and saving resources for the hospital. Assessing credit risk Financial services companies are using AI to sharpen their ability to predict the creditworthiness of loan applications and accelerate the credit-risk assessment process. These capabilities can be key to reducing the losses that come with loans that go into default, according to McKinsey & Company. "With machine learning and other technologies, risk models can become more predictive, which suggests that credit losses may fall by up to 10 percent," McKinsey notes. Even better, a McKinsey survey indicates that over half of risk managers expect credit decision times to fall by 25 percent to 50 percent with the power of AI on the backend. "With machine learning and other technologies, risk models can become more predictive, which suggests that credit losses may fall by up to 10 percent."