Last November, Intel teamed up with the University of Tel Aviv and the Israel Innovation Authority to host AI Week, where 3,000 technologists, researchers, and data scientists from 29 countries gathered to discuss AI innovation across a multitude of industries. Intel was in good company alongside AI hyperscalers like Microsoft, Google, and Amazon, as well as other organizations making strides with AI including Toyota, General Motors, Accenture, Intuit, and Samsung. Joining the 150 attending companies were 16 experts from top universities around the world who spoke at a research symposium led by Intel IT’s Advanced Analytics team.
Through multiple keynotes and sessions, Intel offered a look at some of its latest AI successes, including efficiencies it has facilitated within its own business. From Mobileye to IT to the AI Platforms Group, Intel experts delivered a diverse set of talks on AI research, algorithmic development, industries like healthcare and automotive, hardware, computer vision, and more.
Here are some of the highlights:
- AI Inside: How Intel IT’s Advanced Analytics is Transforming Critical Work at Intel For the last 10 years, the 200 data scientists and data engineers from Intel’s Advanced Analytics team have produced AI-based solutions to transform Intel products, design, manufacturing, validations, and sales.Through failure and triumph, we’ve learned how to execute end-to-end AI solutions that have a real impact on Intel’s business. In this session, Nufar Gaspar, Intel IT Advanced Analytics Vertical Manager, discusses how to identify a good opportunity, the key components for a successful solution, and when to call it quits on something that just isn’t working.
- Using AI Platforms to Accelerate Enterprise AI Deployment What happens after an AI model is deployed? Moty Fania, Principal Engineer for Intel IT Advanced Analytics, explains that machine learning models degrade over time due to data-drift, changing environments, and other factors. Unless models are maintained, their benefits diminish, and they’re likely to break. On the flip side, providing ongoing support for deployed solutions can weigh heavily on resources and distract from the ability to solve new problems. Instead of building many different custom AI applications, Intel’s approach is to build AI platforms capable of rapidly deploying numerous models while incorporating sustainability measures like health indicators, logs, and the ability to re-train models.
- Improving Parkinson’s Disease Research and Care Using AI Parkinson’s disease affects millions of people worldwide. Chen Admati, IT Advanced Analytics Vertical Manager, discusses how her team is using big data and machine learning to advance Parkinson’s disease research by improving the quality of the data being used in clinical trials. Current methods for measuring and tracking the disease’s progression rely on feedback from patients and subjective observations from clinicians. Using sensors that stream information to the cloud while the patient is at home, researchers can use AI to more reliably track symptoms and disease progression, while enabling more effective testing of new drugs.
- Deep Learning-Based Video Inspection for Graphics Processor Testing Dr. Amitai Armon, Intel Chief Data Scientist, describes how the Advanced Analytics team trained a computer model to detect corruptions in videos output by the integrated GPUs inside Intel processors. Using computer vision and a variety of different modeling methods, they reduced human review times by 90% and improved bug-catching accuracy.
© 2020 Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.