Home General AI Frameworks Compete with TensorFlow by Shifting to Interoperability

AI Frameworks Compete with TensorFlow by Shifting to Interoperability

Although artificial intelligence is still in its infancy stages of development, it is likely to touch almost all consumer and business operations in the next 15 years. One of the main aspects or features of artificial intelligence (AI) development is the array of software frameworks required to make it possible.

In this case, the framework entails a collection of interfaces, tools, and libraries that are created for producing AI models including neural networks, which facilitate Deep Learning.

Even with the onset of a considerable rationalization of frameworks, it is highly unlikely for a de-facto framework to come up, especially with the numerous artificial intelligence (AI) use cases that a single framework would require to handle.

ABI Research, a well-known market-foresight consultation firm that offers strategic guidance, particularly on the most innovative technologies recently recognized and benchmarked the main software frameworks that would possibly make up the backbone of potentially any AI product development. The firm did so in an attempt to obtain additional insight into artificial intelligence development.

Even though TensorFlow is the leading software framework according to ABI Research, other numerous frameworks are coming up as potential competitors that are likely to develop various applications and use cases.

These contender frameworks currently do not have the scale and size of developer community interaction, but then again they have been carrying out extreme rationalization, especially around interoperability, in a considerable move to take down TensorFlow’s dominance.

According to Jack Vernon, ABI Research‘s industry analyst the effort by other frameworks is ideal for a flourishing artificial intelligence (AI) technology ecosystem. He added that no single framework could adequately serve all artificial intelligence use cases. As such, he concluded by saying that a diversity of interoperable frameworks would help developers not only in bettering research but also in productizing artificial intelligence.

Currently, the AI technology market is flooded with proprietary technologies. This situation has caused a considerable fragmentation of the artificial intelligence apps development as well as creating confusion for both implementers and developers of AI technology. As seen previously in other industries like Internet browsers, desktop operating systems, and mobile devices ecosystems, technology rationalization is the main milestone of technology development. Hence, artificial intelligence is no exception.

Frameworks are witnessing rationalization around two main factors. Firstly, some frameworks have been upgraded to accommodate a great scale of deep learning methods, or they have died off in regards to community and developer support.

On the other hand, there have been some frameworks that govern bodies, which have chosen to work together around a series of principles that allow deep learning (DL) models to be interchanged between them. For instance, Vernon explained that the Open Neural Network Exchange would rebalance the framework space, which is excessively reliant on the success of Tensorflow and its supporter, Google.

ABI Research has evaluated the AI frameworks ecosystem and created a benchmark that measures several KPIs including hardware portability, scalability, developer interest, future-proofing, governance, reliability, hardware efficiency, and edge accessibility among others. This first of its kind report found that Tensorflow was the leading framework followed by others like Caffe2 and MXNet in that order.

Source PRNewswire

Subscribe to our newsletter

Signup today for free and be the first to get notified on the latest news and insights on artificial intelligence

KC Cheung
KC Cheung has over 18 years experience in the technology industry including media, payments, and software and has a keen interest in artificial intelligence, machine learning, deep learning, neural networks and its applications in business. Over the years he has worked with some of the leading technology companies, building and growing dynamic teams in a fast moving international environment.
- Advertisment -


AI Model Development isn’t the End; it’s the Beginning

AI model development isn’t the end; it’s the beginning. Like children, successful models need continuous nurturing and monitoring throughout their lifecycle. Parenting is exhilarating and, if...