What you need to know about AI Computing at the Edge

I recently wrote about Artificial Intelligence (AI) and edge computing in an article for Forbes. Today, more and more data is being collected by a growing number of IoT devices and there is growing interest in edge computing. What does that mean for the enterprise, and what does the future hold for AI at the edge?

There’s room for both the cloud and the edge

In the article, I discussed how edge computing will continue to co-exist with cloud computing. The fact is, cloud computing is not going away anytime soon and there’s a time and place for both computing in the cloud and at the edge. Cloud computing has become the gold standard for enterprise computing today because it solves the problem of storing and managing all types of applications, as well as the huge volumes of data required to develop and maintain algorithms for AI apps.

At the same time, there is a growing interest in enterprise edge computing, since it makes sense to develop AI functionality closer to a data source, especially when you may need to know, and possibly act on, information in real time – and when a situation may change rapidly. For example, consider a video surveillance application where you need to be able to identify if someone is acting suspiciously or aggressively, or predict dangerous behavior based on a person’s actions.

The challenges of AI at the edge

The article also explored the challenges involved in developing AI algorithms for edge computing. In addition to the typical issues of the shortage of data scientists to program the apps and the specialized GPU servers needed to crunch the data, the limitations of edge computing make it even more difficult for AI development. IoT and other edge devices currently lack the processing power and bandwidth that is needed for this process.

So, while there may be interest in AI at the edge, there is still not a feasible way to develop it – right now, that is. Many companies are busy working on this problem and testing new solutions that are on the horizon. Yet there’s no need to wait and put your AI plans on hold — there are a lot of business problems that you can solve right now using cloud AI to move your business forward.

For additional AI insights from Wovenware, you can read: Are You Prepared for AI in 2019? Key Trends Driving its Growth and AI is All Around Us – Sometimes Without Us Even Noticing It.

AI Can Take the Med Device Market to the Next Level

In the past few months I’ve spoken at a medical device conference, and shared insights in industry publications – most recently in Medical Design and Outsourcing, about the role of Artificial Intelligence (AI) in med device manufacturing and supply chain operations. The market is prime for AI, given the innovation taking place in medical devices and the critical role they play in patient health.

The Medical Design and Outsourcing article, which included the thoughts of many industry experts, including Todd Morley, director of data science for Medtronics, Prashant Jagarlapudi of Sparta Systems and others, focused on the fact that AI can play a significant role in the efficiency of medical device manufacturing and reduce risk, but they caution that it’s still a work in progress.

Medtronic’s Todd Morley told Medical Design & Outsourcing that his company anticipates widespread use of AI. He said, “Industrial engineers have applied statistical methods to manufacturing for decades, however, the convergence of ubiquitous, inexpensive sensors; abundant computing resources; and powerful, highly accurate AI methods such as deep learning and graphical modeling creates new business cases for AI in manufacturing.”

In the article I shared the idea that through the support of AI, engineers can have more time to innovate, instead of fixing problems. For example, a defective device that may be taken out of production could use a predictive algorithm to determine the probability of that particular device being scrapped, so it won’t be sent to the engineering department.

While the use of AI in med device manufacturing may still be in the early stages, it will soon be a core part of the process. As I mentioned in the article, just as AI is helping with autonomous driving or suggesting which movies we should stream, it could become so omnipresent in med device manufacturing that “people won’t even notice it’s there.”

Why Outsourcing to Puerto Rico Is a Shore Thing

An article I wrote recently for The Future of Sourcing summed up the reasons why nearshoring to Puerto Rico brings the best of outsourcing to U.S. businesses – the cost benefits of offshoring and the productivity benefits of nearshoring.

And, proof is in the adoption. U.S. companies are quickly realizing the benefits, with business process outsourcing (BPO) becoming a growth industry for the island.

The article examined the key benefits of outsourcing to Puerto Rico, such as the following:

U.S. commonality.

As a U.S. territory, Puerto Rico shares the same currency, similar time zones, languages, intellectual Property (IP) protection and other U.S. regulation adherence, conveniences that would be impossible to replicate with offshore providers, but at billing rates still 30-50 percent cheaper than mainland competitors.

Closeness for collaboration.

As software development continues to become a strategic business driver, it’s important for to companies to be able to easily collaborate on-site with outsourcing partners. As I said in the article, “today’s outsourced provider must be a true business partner that understands your business – not some obscure firm with generic answers.”

Technology expertise.

Puerto Rico home to world-class universities and many of its professionals also attended leading American universities on the mainland. Their tech expertise is on par with all major U.S. cities.

 

Nearshoring to Puerto Rico has many advantages that are making it an attractive destination for AI and software development. As the complexity of tech grows, along with the need to bring advanced solutions to life quickly and efficiently, so too will the market demand.

For additional nearshoring insights from Wovenware, please read: Why it Pays to Outsource Closer to Home and Nearshoring: The Fast Path and Short Cut to AI Deployment.

AI is All Around Us – Sometimes Without Us Even Noticing It

In a recent Forbes article, I discussed the ubiquity of AI, and how it is seeping into just about every aspect of our lives – for the better.

AI has potential to change the world in major ways — helping to cure cancer, improve the environment and generally make our lives easier and safer. But more pragmatic uses of AI are already changing our lives today.

How is AI at work in our everyday lives? Below are a few highlights from the article:

  • Smart spam filters. What person on the planet doesn’t feel like he’s drowning in emails, but did you know that without AI you would be getting a whole lot more junk mail? Most Internet Service Providers (ISPs) use AI algorithms to filter them out before they even reach our inbox – by looking for patterns in the emails that might point to spam.
  • Chatbots and virtual assistants. This form of chatty AI is popping up everywhere, and many times we don’t even realize that we are interacting with it. While some chatbots are still working on perfecting speech recognition and have a ways to go, they are continually learning and getting smarter all the time.
  • Navigation systems. We rely daily on map and navigation apps to get us where we’re going, helping us re-route in real time if there’s a slow down or accident up ahead, or alerting us to speed traps when we’re driving too fast. These navigations systems use complex algorithms to do all of this, leveraging huge datasets that are constantly being updated.
  • Personalized marketing. Chances are you have received personalized recommendations based on your buying or viewing history. While this can be annoying at times, sometimes it’s nice to know that someone understands you.

The Forbes article only shares a few examples of pervasive AI, but the list could go on and on. Take a look around and you may be surprised at how many everyday activities and objects are powered by AI in some form — and it’s only the tip of the iceberg.

Healthcare and EDI: A Match Made in Data Heaven

Electronic Data Interchange (EDI), has been around for decades.  It is a process through which data is exchanged in a standardized and structured format so that different data systems can contain the same information almost instantly. A good comparison might be the way a person speaking English might communicate with a person speaking Russian. They often need a middleman to serve as a translator for both. That’s what EDI does for systems – serves as the middleman so that system A can synchronize data with system B seamlessly.

Several standards exist to format the interchangeable data between systems. Some major sets of EDI standards, include:

  • The UN-recommended UN/EDIFACT, the only international standard, predominant outside of North America
  • The US standard ANSI ASC X12 (X12), which is predominant in North America
  • GS1 – An EDI set of standards predominant in global supply chains
  • The TRADACOMS standard,developed by the ANA (Article Number Association now known as GS1 UK).  It is ominant in the UK retail industry
  • The ODETTE standard, used within the European automotive industry
  • The VDA standard, used within the European automotive industry mainly in Germany
  • The HL7, a semantic interoperability standard used for healthcare administrative data.

We use EDI in many industries where standard data exchange is needed, and a key industry is healthcare. It is critical for the healthcare industry to adopt EDI-enabled systems, such as ONC Certified EHR systems and HIPAA X12 compliant billing systems, since the exchange of data between providers, payers and patients is critical. The adoption of these systems has a direct impact on all parties involved.

Here is a simple example of how EDI can help your visit to the doctor be more rewarding:

Imagine if you went to the cardiologist for the first time because of a recently diagnosed condition – a condition for which you have visited your primary care provider a couple of times. Your social history, demographic and medical history data will be asked by the cardiologist.  However, that information already exists in the primary care provider’s Electronic Health Record (EHR) system.  Using EDI, your cardiologist could receive this information in just a few seconds, enabling him to better focus on your condition, instead of wasting time asking questions you already answered somewhere else. And, since a good doctor cannot make a good diagnosis without all the necessary facts, the doctor can rest assured that he is receiving accurate information

The main goal of EDI in healthcare is to provide up-to-date information on a patient’s condition across the healthcare ecosystem. It helps reduce common mistakes that often occur when handling patient information and helps reduce fraudulent behavior by both the providers and the payers.

As an EDI senior software developer with over 11 years of experience, I have seen first-hand the complexity that is involved with EDI and all the challenges it still has yet to address.  But what if Artificial Intelligence (AI) was included in the mix, helping EDI developers create better and more efficient systems?  This could lead to better and faster communication between systems and go beyond the boundaries of formats and standards.  AI could help improve the way healthcare systems communicate with each other and fill in the gaps when additional information may be required.

As an example, take the interoperability standard, HL7, which contains two different versions that both can be used at the same time.  The two versions allow for the transmission of different types of data.  With the proper AI implementation, a developer could quickly close the gap between both formats by relying on the AI capabilities to determine the required output regardless of the versions and then just validating that the output is the expected one.  This would significantly reduce development time along the way.

EDI has transformed the way data is shared in healthcare for improved patient outcomes, streamlined processes and reduced healthcare costs. Making it smarter through the use of AI can go a long way to putting these benefits into overdrive.

Can Artificial Intelligence (AI) Really Be Used to Help Heal the Planet?

An article I wrote recently for Forbes discusses the transition of Artificial Intelligence (AI) from an enterprise app to one being used in the field — quite literally.

AI has been known for helping businesses make better decisions, automate core tasks and other activities, but it also is playing a critical role helping to protect our planet, preserving wildlife, preventing forest fires, protecting marine life, and in many other ways that may not get the same attention as business apps.

As I’ve discussed recently, here in Puerto Rico Wovenware has been fortunate to participate in an important environmental and health project to help prevent mosquito-borne illnesses.

It’s very rewarding to apply our AI expertise to help address issues impacting the general population and environment. But the mosquito project is just one example of environmental AI; there are so many more.

Machine learning is being used to monitor wildlife and determine the population of snow leopards. It’s also being used to predict where forest fires might erupt, by using historical meteorological data and an artificial neural network; helping to combat wildlife poachers; and using satellite imagery to track coral bleaching in the oceans and look for marine life diseases and pollution.

As the Forbes article explains, by augmenting the time-consuming, manual counting and categorizing activities typically done by humans to create massive datasets, AI is letting us go to places that were previously impossible. AI is no longer about helping businesses succeed, but healing our common planet as well.

How else do you see AI being applied out in the world?