By: Anders Lentell

2019-12-17

What do you have in mind when you think AI 2020?

“Just as electricity transformed almost everything 100 years ago, today I actually have a hard time thinking of an industry that I don’t think Artificial Intelligence and Machine Learning will transform in the next several years” – Andrew Ng

Machine learning algorithms are already becoming a natural part of systems within healthcare, automobile, manufacturing, entertainment, agriculture and more. While most of us still look at AI as something out of a sci-fi movie where robots are taking over, the front line of developers are already implementing new and efficient algorithms improving performance and making systems more adaptable. So where is this going? What are the trends that will dominate the next decade? Some things to keep an eye on 2020 are:

1: Internet of Things <3 Machine Learning

Machine Learning feeds on data, IoT devices generate lots of data and relies on intelligent responses to take actions. It is a perfect match, and this is a tech explosion we haven’t even begun to see yet. According to Business Insider, there will be more than 64 billion IoT devices by 2025, up from about 10 billion in 2018. The need for advanced deep learning models will grow with the explosion of all the new connected technology that is coming.
This convergence of IoT and ML can transform industries and help them in making more informed decisions based on the huge amounts of data available every day which will result in new value propositions, business models, revenue streams and services.
On top of this we see advances in edge computing enabling even better interaction between IoT devices and deep learning models. This opens for autonomous systems and data qualification already at the source making even faster and more precise implementations.
Edge computing and AI also opens for de-centralization of AI which will be the biggest driver for enterprise AI structures and strategies in 2020

2: “Are you talking to me?”

To have a “talking” system has been in focus for a while and we expect to be greeted by an AI powered chatbot in most online customer centres already. Conversational AI or Natural Language Processing has been developed for some time but is expected to really bloom in 2020. Next year, Gartner predicts that 50 % of analytical queries will come from NLP systems and that 70% of all white-collar workers will work with conversational platforms every day.

3: Digital flooding

“Data is the new oil” we keep hearing and in many ways it is true. Producing and maintaining quality data is increasingly important and having a good data set is something you can sell and profit from but at the same time we produce data in an exponentially growing pace and keeping these sets are challenging in sheer hardware and we also see risks with integrity and personal privacy. Building models that learn so that we can keep the conclusions and drop the data will be a new necessity. This will make it even more important to analyse and identify what to keep and what to clean out. In parallel to this we hear a growing demand for security protocols and safety systems to ensure that our prized data is kept under control and untainted.

4: Automated Machine Learning and high-level tools

While the black magic of Machine Learning might be difficult and hard to grasp and master there are more and more tools available for easy training of even complex models without needing to dig deep into how the algorithms work.
This will accelerate the use of AI and open for new use cases that can benefit from implementing machine learning. Even if these high-level tools will help many take the plunge into AI and in many cases be enough to make powerful solutions it is not a silver bullet for all and the need to build more advanced models will still be there.

Whether your goal is to carefully move into the AI community by implementing your first learning algorithms or you already are well on your way in transforming your company to be AI driven, maybe even having AI making executive decisions, choosing the right approach for the task at hand is crucial.
The options are almost endless and often companies tend to overspend in the quest for new technology. Classic machine learning will still be important. Gartner predicts that in 2022 more than 75% of all companies will be implementing deep neural networks where classic machine learning would have sufficed. The importance in understanding the different machine learning options, keeping things simple and choosing the appropriate technique should be a priority to ensure you get most bang for the buck!

#AI #Internet of things #Machine learning

Industrial Internet of Things

Att lära sig Kotlin