Artificial intelligence is getting more and more accessible. As a result, AI-based developments acquire a local character tapping into standalone companies. At the same time, technologies are transforming, blending, and merging to create something more large-scale. Read further in the article what to expect from the artificial intelligence field in 2019.
AI specialists will be trained, not looked for
In the light of the technological race, the war for IT talents is getting more and more severe, especially, taking into account that the demand for AI developers exceeds supply.
The best AI specialists can earn millions of dollars now, thanks to China in particular. The salary of a Chinese senior scientist that studies artificial intelligence equals to $567-624 thousand per year. Experts in machine learning from other countries make $567-624 thousand for the same period.
Chief Technology Officer of LexisNexis computer service provider Jeff Reihl says that forward-thinking IT leaders will not wait for the wanted developers to come to the labor market. In 2019, they will be investing in the development of required skills in existing teams.
Reihl says, “Companies will take a more serious approach to educating teams about such advanced technologies as AI and ML.”
AI is a helping hand, not a substitute, and it will become obvious
One of the most notorious and exciting news about AI is that the technology will force people out of workplaces. There are reasons for such thoughts, as artificial intelligence and related technologies may influence workplaces in the future. However, today AI is more likely to facilitate the growth of teams rather than reduction.
CEO of Unsupervised Noah Horton believes, “The trend of 2019 is the implementation of AI solutions that will allow teams with restricted resources to work more productively.”
In fact, IT companies do not want AI to replace people. They hope that the technology will help to solve critical problems fast, especially when the personnel is scarce.
“AI is not a magic wand for overcoming the gap in skills, but it is important for solving some difficult and resource-intensive tasks.” CEO of Awake Security, developer of cyber security systems, Rahul Kashyap notes, “I think companies that will maintain the balance between people and machines – both within enterprises and as part of their customer-oriented products – will achieve the greatest success with AI next year.”
Convergence of IoT and AI will take place
Industrial IoT may become the biggest driver of implementing AI at enterprises. Together two technologies will be able to analyze the state of hardware and carry out maintenance service, allowing businesses to save on repairs significantly.
For this purpose, advanced ML models based on deep neural networks are optimized to work with IoT. Such neural networks will process data generated by cameras, microphones, and other sensors.
Head of IoT at Technoserv Andrey Shuravin thinks that IIoTprojects with artificial intelligence will be relatively inexpensive and quickly realizable, and will allow receiving results almost immediately.
In the interview for Tadviser, Shuravin said, “Adoption of IIoT allows reducing expenditures related to equipment operation, stock management, and industrial safety. Interest in such projects at production sites will lead to the development of the market.”
Chatbots will help with chores
Alexa and Yandex Alice virtual assistants became highly popular among users in 2018. According to data from Yandex, Alice could boast monthly audience of 30 million users. So, one can expect even more speech recognition tools in 2019.
Besides, producers will start integrating chatbots in home appliances. In December of 2018, Sony, TiVo and Hisense presented voice controlled TV sets. Such appliance producers as Delta, Whirlpool and LG added Alexato their devices to help people control everything in their houses, from microwave ovens to faucets.
Different neural networks will learn to interact with one another
To build neural networks, developers have to select one tool from the variety of options, for example, Caffe2, PyTorch, Apache MXNet, Microsoft Cognitive Toolkit, and TensorFlow. However, after a model is trained and assessed in a specific structure, it is hardly possible to shift it to another structure.
Previously, there was no interaction between tools for neural systems development and it hindered their development. In 2017, AWS, Facebook, and Microsoft joined efforts to develop Open Neural Network Exchange (ONNX), which allows reusing trained models of neural networks in different structures. At the beginning of December 2018, Microsoft revealed source codes of ONNXexecutive environment.
In 2019, ONNX will become an important technology for the industry. Thanks to ONNX and other Azure AI services, developers and scientists will find it easier to make new inventions in the field of artificial intelligence.