Enterprise Services, Service Provider Markets
By Camille Mendler 01 Sep 2020
Enterprise views of private 5G networks and their suppliers; demand trends post-COVID 19
Deep learning (DL) has dramatically improved the capability of artificial intelligence (AI) systems in recent years. We expect to see two main developments in 2017. First is the wider application of AI systems in various domains, such as healthcare, agriculture, telecommunications, retail, and finance, as well as in combination with emergent technologies such as Internet of Things (IoT) systems and analytics for big data. Second are the new hardware accelerators due to appear in 2017 that are likely to further improve the algorithms. The impact that AI will have on society is beginning to be addressed. 2016 saw a lot of activity in the US at federal level and it is likely that similar initiatives will begin in other countries.
DL neural networks within the machine learning branch of AI represent the most successful innovation yet achieved in the field of AI. The moment in March 2016 when Google DeepMind’s AlphaGo machine (based on DL) beat world Go champion Lee Sedol four games to one was a milestone in human history. This AI technology will permeate many application areas in 2017, ranging from autonomous driving to a wide variety of Internet of Things applications, from consumer products to healthcare. IoT in particular will generate big data too vast for humans to process and AI will play a major role in analyzing and making sense of streaming data and content in data lakes.
Enterprises and vendors alike need to address the implications of AI in their sector in 2017 to meet the challenges and opportunities that arise from what is likely to be the largest and most profound technology wave yet. Putting together a beginning strategy for AI should be on every organization’s agenda for 2017.
The key technological innovation today that DL introduces is powerful self-learning. This means that the designer and developer of the intelligent machine does not need to be an expert in the application domain when building the AI system because the AI system learns to acquire new skills through the DL training algorithms. These AI systems can also continuously learn and improve when used in real work scenarios. A good example is AI-powered chatbots that are also assisted by humans. The chatbots are learning from human intervention, and one can surmise that one day chatbots will run autonomously.