On Tuesday, Chris Reece looked at the impact of artificial intelligence on the telecom industry at the recently held Mobile World Congress. The Award-Winning technologist explained that the AI that is currently used by the telecommunications industry could assist in solving some of the most complex issues for communications service providers (CSPs).
According to Reece, CSPs have adopted artificial intelligence gradually since the initial issues AI was intend to solve did not significantly affect their operations. This situation made him ask the crowd for some examples of matters solved by AI. With answers like image recognition and chess, Reece acknowledged that he did not know many telecom companies that require a computer to identify the difference between a dog and a cat or vice versa.
Reece added that the telecom space hosted numerous opportunities for AI use. He also hinted that the industry was beginning to adopt the breakthrough technology, which would eventually help it solve some of the industry’s leading issues like network planning and fraud management.
The good thing about AI is that it can use machine learning to train itself to solve better the issue it was made to eradicate initially. Reece explained this capability by drawing his argument from a hypothetical AI used in differentiating cat images from those of dogs. In this case, he said that it is possible to come up with an algorithm that can easily distinguish them. However, machine learning can help you come up with a more powerful AI.
AI is provided with outputs coupled with inputs, especially in machine learning. In this case, the hypothetical AI would be fed with cat and dog images as inputs. In turn, the AI can analyze the details that the cat images have in common. The same case applies to dog images.
For Reece, the telecom industry is currently dealing with low-hanging fruit issues such as automating customer billing. He also hinted that most of the companies that demonstrated machine learning at the MWC mainly looked at identifying new ways of monetizing subscriber data. Additionally, Reece acknowledged Peter-Service, a St. Petersburg-based company that has created an AI with the ability to target distinct consumer demographics using digital billboards.
Arm, a device architecture firm, also exhibited its Project Trillium, which entails a mobile AI project. The company has created its software library dubbed Arm NN in a bid to train AIs. Arm also boasts programmable processors that are based on mobile phone neural networks. These processors provide devices with machine learning abilities without the need for Internet connection, more precisely for extra computational power.
Aside from such innovations, Reece predicts bigger things in the future that would solve issues like network planning or fraud management in the telecommunication industry. For him, a company with the potential to accurately spot and independently manage fraud cases would have a greater advantage over its counterparts. Although no company showed such AI potential during MWC, DARPA’s advancements in machine learning serve as a beacon of hope for the telecom industry.