The original link

4 reasons why the next wave of mobile apps will be powered by machine learning.

| the author Karl Utermohlen

Ai/machine | link heartbeat. Fritz. – lea…

Mobile developers can benefit from the revolutionary changes on-device machine learning can offer. This is because the technology can support mobile applications, allowing for a smoother user experience by leveraging powerful features such as accurate location-based recommendations or instant detection of plant diseases.

This rapid development of mobile machine learning has been a response to many of the common problems faced by classical Machine learning. In fact, they are about to happen. Future mobile applications will require faster processing speeds and lower latency.

You might wonder why AI-first mobile applications can’t simply reason in the cloud. First, cloud technology relies on a central node (think of a large data center with lots of storage space and computing power). This centralized approach simply cannot meet the processing speed required to create a fluid, machine-learning-driven mobile user experience. Data must be processed in this centralized data center and then sent back to the device. This takes time and money, and it is difficult to keep the data private.

Having outlined these core advantages of mobile machine learning, let’s explore in more detail why, as a mobile application developer, you’ll want to stay tuned for the upcoming device machine learning revolution.

Reduce the delay

Mobile App developers know that high latency is a major cause of failure for an App, no matter how powerful it is or how reputable its brand is. Many video apps for Android devices have had latency issues in the past, leading to an out-of-sync experience when viewing audio and video. Similarly, a social app with high latency can lead to a very frustrating poor user experience.

Because of these latency issues, machine learning on mobile devices is becoming increasingly important. Consider social media image filters and location-based meal suggestions — these app features require low latency to deliver the highest level of results.

As mentioned earlier, cloud processing times can be slow, and ultimately developers need to reach zero latency for machine learning to work in their mobile applications. Machine learning on the device paves the way for near-zero latency through its data-processing capabilities.

Here is an example of real time low latency: style conversion results for live video in a Heartbeat application.

Smartphone makers and big tech companies are catching up. Apple, which has been leading the way, is developing more advanced smartphone chips using its Bionic system, which has a complete neural engine that helps neural networks run directly on the device with incredible processing speeds.

Apple continues to iterate on Core ML, a machine learning platform for mobile developers; TensorFlow Lite adds GPU support; Google continues to add preloading features to its own machine learning platform, ML Kit. These are among the technologies that mobile developers use to create applications that can process data at lightning speed, eliminate latency, and reduce errors.

This combination of accuracy and a seamless user experience is the number one consideration for mobile developers when creating ML-driven applications. To ensure this, developers need to embrace and embrace machine learning on devices.

Enhance security and privacy

Another great advantage of Edge computing that should not be underestimated is how it improves security and privacy for its users. Ensuring the Protection and privacy of application Data is an integral part of mobile developers’ work, especially given the need to meet the General Data Protection Regulations (GDPR), these new privacy laws will certainly affect mobile development practices.

Because the data does not need to be sent to a server or the cloud for processing, cybercriminals have little opportunity to exploit any vulnerabilities in the data transmission, thus ensuring that the data is not compromised. This makes it easier for mobile developers to meet the GDPR’s data security requirements.

Machine learning solutions on devices also offer decentralization, much like blockchain does. In other words, a DDOS attack makes it harder for a hacker to destroy the network connection of a hidden device than the same attack on a centralized server. The technology could also prove useful for drones and future law enforcement efforts.

Apple’s smartphone chips, which are the backbone of Face ID, for example, could also help improve user security and privacy. The iPhone feature relies on a neural network on the device to collect data from all the different dimensions of a user’s face as a more accurate and secure method of identification.

  • Apple introduces Face ID on iPhone X link to video: www.youtube.com/watch?v=z-t…

This and future AI hardware will pave the way for a more secure smartphone experience for users and provide mobile developers with additional layers of encryption to protect users’ data.

No network connection required

In addition to latency issues, sending data to the cloud for inferential calculations requires a valid Internet connection. Usually, in the more developed parts of the world, this approach can be easily implemented. But what about places with poor Internet connectivity? With machine learning on the device, the neural network can run directly on the phone. This allows developers to use the technology at any given time and on any device, regardless of network connectivity. In addition, it can democratize machine learning features because users do not need an Internet connection to their applications.

Healthcare is an industry that could benefit greatly from machine learning on devices, as app developers are able to create medical tools to check vital signs and even perform remote robotic surgery without any Internet connection. The technology could also help students who need to access classroom materials in places where there is no Internet connection, such as in public transport tunnels.

Machine learning on devices will eventually give mobile developers the tools to create applications that can benefit users around the world, regardless of their Internet connection. Even without an Internet connection, the new smartphones of the future will be so powerful that users will not suffer from latency problems when using apps in offline environments.

Reduce business overhead costs

Machine learning on your device can also save you money because you don’t have to pay external vendors to implement or maintain these solutions. As mentioned earlier, you don’t need cloud computing or the Internet to provide such solutions.

Gpus and AI dedicated chips will be the most expensive cloud services you can buy. Running the model on the device means you don’t have to pay for these clusters, thanks to the increasingly complex Neural Processing Units (NPUS) found in today’s smartphones.

Avoiding the onerous data-processing nightmare between mobile and cloud is a huge cost savings for businesses that opt for machine learning solutions on devices. Through on-device inference on such equipment, bandwidth demand can also be reduced and a great deal of cost can be saved eventually.

Mobile developers can also significantly save money on the development process because they don’t have to build and maintain additional cloud infrastructure. Instead, they can achieve more with a smaller engineering team, which allows them to expand their development team more effectively.

conclusion

There’s no doubt that cloud computing has been a boon for data and computing in the 2010s, but the tech industry is moving at an exponential rate and on-device machine learning on devices may soon become the standard for mobile apps and iot development.

With its lower latency, enhanced security, offline capabilities and reduced costs, there’s no doubt that all the major players in the industry are taking a big look at this technology, which will define how mobile developers can advance the creation of applications.

If you’re interested in learning more about mobile machine learning, how it works, and why it’s so important in the overall mobile development space, here are some additional resources to help you get started:

  • Matthijs Holleman’s blog Machine, Think! There are a lot of great tutorials and other content on Apple’s mobile machine learning framework, Core ML;
  • Edge Artificial Intelligence (Video)
  • Heartbeat, of course, has a growing library of resources at the intersection of mobile development and machine learning.

Pay attention to our

Please follow our official account ios-tips and join our group to discuss issues. Add coldlight_HH/wSY9871 to our iOS/ FLUTTER wechat group.