In the last few years, scientists and technologists have made great strides in the field of Artificial intelligence: machines have beaten humans at chess and Go, there’s now a robot that can navigate without GPS, China has taken facial recognition to a whole new level with its adoption of machine learning in daily policing, not to mention, Google is still wowing the world with its self-driving cars.
2020 is likely to see more growth, more experiments and great strides towards more adoption of AI, as governments, scientific bodies and tech giants like Amazon, Facebook, and Google continue to throw more resources into research and development in AI fields such as robotics, NLP, computer vision, and Machine Learning.
PWC’s money tree report shows an increase in funding by close to $4billion for AI startups, from 2017 to 2018.
Figure 1: AI funding increased by $4bn in 2018
There are also quite a number of new entrants in the field of Artificial Intelligence, both from tech and non-tech backgrounds and from predictions, these new companies are the ones likely to break the trend of tech giants and take AI more mainstream.
So, what kind of advancements should you expect to see in AI? For more on this, read below.
Cloud-based AI will increase AI adoption by enterprises
Research by Deloitte Global shows that 70% of enterprises will become AI-enabled by adopting cloud-based software. Whereas AI has been a preserve of a few tech giants, cloud-based AI will increase the adoption of AI by enterprises and better return on investment.
Cloud-based AI also solves the skills-problem that has stalled progress in the industry. After surveying 1900 experts from 10 industries, Deloitte found that the “skills gap” is one of the main obstacles against AI adoption. 41% of the respondents said they had a “moderate” skills gap while 27% said their skills gap was extreme. With cloud-based software, companies of all sizes can take advantage of AI capabilities regardless of skill level or funds to invest in research. In fact, respondents are in favor of using enterprise that is cloud-based and AI-enabled, as opposed to investing funds to research and build their own software.
Figure 2: Companies favor cloud AI as opposed to building their own. (pg17)
Key players will take an interest in developing their own AI technology
AI requires powerful processors to work, but one drawback to date is that though chipset manufacturers like Intel, Nvidia, AMD, Qualcomm, and ARM have made great strides in improving the specs of their processors, there is always a feature that adopters of AI must contend without.
This has led companies like Amazon, Facebook and Google to explore the option of building chips that they can customize to their huge data needs. Facebook, for instance, is backing Intel’s efforts to build its new processor, Neural Network Processor for Inference (NNP-I). Intel made the announcement at CES 2019. Google has also built Edge Tensor processing units (TPUs), an AI invention that will enable machine learning on IoT devices.
Better chip technology will enhance the speed of AI applications, increase machine learning ability and more use cases for specialized tasks that require extensive calculations and inference.
Computer vision continues to grow
Computer vision enables computers to “see” one or more images, understand it and extract meaning from it. Demand for computer vision will increase in 2020, as interest in augmented and virtual reality, drones, autonomous and semi-autonomous vehicles continues to grow. Research and development will go to the creation of advanced cameras and images sensors, in addition to more sophisticated deep learning software applications.
More so there will be,
● More application of computer vision for quality control in the pharmaceutical, automotive, medical and food industries.
● The growth of deep learning algorithms that allow speedier and more accurate detection and classification of images.
However, there are still concerns with security, particularly with the development of self-driving cars. Research and development will continue to increase accuracy and develop low-cost affordable hardware.
Robotics will enter consumer markets
2020 will see more industrial automation as companies seek to streamline processes and cut costs. Company heads will increasingly free their workers from routine tasks or tasks that require machine-like precision. This will result in more streamlined workforces, more job satisfaction, and quality assurance.
Experts also note that robots will also enter the consumer space. In 2016, the president of the International Federation of Robotics, Joe Gemma, predicted that by 2019, sales of service robots would have grown to $22 billion. That prediction seems to have held true as service robots have since been created to assist in functions such as sleep (Somnox) and cleaning (iRobot).
Julian Jagtenberg, creator of the Somnox sleep robot says that we are likely to see a rise in robots that are safer for humans to interact with. He cites healthcare as one industry that could benefit from robotic companions that can assist patients in daily tasks such as bathing and taking medication. This will free doctors and nurses to focus on patient recovery.
Chris Jones, the CTO for iRobot, the creators of the cleaning robot, Roomba, predicts three main advancements in robotics for the home sector:
● Creation of devices that have more spatial awareness. For instance, cleaning robots that can identify areas of your home.
● A smart-device ecosystem with devices that can integrate. Devices that cannot integrate will have to be weeded out.
● The “smart home 2.0”, characterized by thoughtful devices capable of customizing task automation based on your environment and preferences.
IDC forecasts also show that demand for devices like smart speakers will continue to rise. For instance, sales of Amazon Echo and Google Home will grow by 39.1% year on year, to reach 1.3 billion devices by 2022. One can learn how to automate the robotic process, through this RPA using UiPath Certification Training Course
Deep learning emerges in Natural Language Processing
Natural Language and machine learning are the most widely used forms of artificial intelligence. In the past, NLP has been most instrumental in text analytics, and this will continue to be the case. However, new applications of NLP will be seen in semantic search and speech to text. At the core of this development is deep learning.
Mary Beth Moore, SAS Artificial Intelligence and Language Analytics Strategist note that the use of Deep learning for natural language processing will lead to higher accuracy in text analytics. Moore also points out that deep learning deployments into speech analytics will transform template-based speech recognition systems such as chatbots into sophisticated applications that deliver a more realistic and less frustrating experience to end-users.
From the above, we can conclude that AI will continue to have a big impact on our lives, and Dr. Santanu Bhattacharya has a few opinions about what this means. For instance, he foretells a continued end to ownership as AI continues to influence the rise of businesses based on subscription models. Companies like Wag, WeWork, Letote, and Common, make it possible for you to subscribe to anything from dog-walking services, office space and even clothes. For the consumer, these advancements will also mean you’ll have an easier time learning, shopping, accessing customer care, accessing better health care, among many other benefits.