Body language plays a key role in effective communication. What messages are you sending? Consider these tips for more impactful and engaging Zoom meetings
Edge computing and AI: 7 things to know
How do edge computing and artificial intelligence (AI) work together? Why does edge fit well with AI? What are some use cases? Let's examine what IT leaders should know
5. Edge AI is data-hungry
This whole process works only if you have enough data to build a statistically relevant model, says IDC’s McCarthy. “Many companies do not meet the minimum requirements, whether it be in terms of volume of historical data or the right kind of data to achieve the desired outcome.”
6. Start by getting your data house in order
Most organizations have not built comprehensive data management practices and do not have these types of data sets on hand, according to CompTIA’s Robinson. “In addition, modern AI is based more on probability than previous software programs. The risk of incorrect or nonsensical answers is higher, and that risk increases if the training data is incomplete or biased in any way. Rather than being able to quickly install an AI component and reap the benefits, companies need to start with a thorough examination of their data.”
In the meantime, subject matter experts can substitute business logic for data-based learning that can be used in real time against multiple data streams, according to McCarthy, until the organization has amassed enough good data to fully take advantage of AI.
7. The cloud-to-edge architecture should be flexible and forward-looking
“As you define the architectures, make sure that you are designing for enterprise scale,” says Mann of SAS. “The cloud-to-edge architecture needs to support deployment of models, model changes over time, and transmission of data in a secure environment.”
Mann advises implementing architectures that are agnostic to chip sets, operating systems, and cloud providers to provide the greatest flexibility for sustained value over time.
While not all problems are fit for edge AI, says Mann, “all IT infrastructure and architectures should be designed to accommodate analytics at the edge as advanced use cases develop. It’s important that you have an environment that can support the deployment of analytics in the required location for real-time or batch processes.”
[ Want to learn more about implementing edge computing? Read the blog: How to implement edge infrastructure in a maintainable and scalable way. ]