IoT version 2.0 – The real Artificial Intelligence
IoT fire is catching up in every industry. It is changing the businesses drastically than we imagined. This blog covers some of the aspects that will fuel the realization of IoT and brings businesses into a new territory. Thing – refers to devices, sensor, software and everything else.
What is version 1.0?
Billions of Things are already connected to the Internet. Currently, most of the effort spent in IoT can be classified into 4 parts.
- Enablement – Making the Thing connected to the Internet
- Business decisions – Analytics based on the data generated.
- Interoperability - Connecting the 'Things' with other 'Things' seamlessly
- Privacy & Security
In the next few years, we expect more and more Things to be connected to the Internet. I believe, the version 1.0 or the first major version of IoT means all (If not, most of) the Things that are connected to the internet will be interoperable. Also, the Privacy & Security concerns are addressed and standards will emerge.
What's next?
All of us know Artificial Intelligence (AI) is going to take the world like a tsunami. So, what is the relation between IoT and AI? AI needs a logical infrastructure to realize its huge potential. IoT version 1.0 can provide that base by enabling all the Things connected and they are interoperable.
Consumer IoT
Consumer based Things are dominating the IoT market in 2015. Predictive analytics, prescriptive analytics, etc. based on Machine learning add tremendous value to the success of the business. These are just the starting points that help consumer based market to make intelligent decisions based on data generated from the Things.
AI adoption on consumer based Things will be slower in comparison. A lot of these Things are of low cost and have low processing power. For Example, a sensor enabled smart device installed in a smart home. AI is not a straightforward implementation in the device. But AI could be a cloud based solution that is taking the intelligent decisions real-time based on the real-time data.
Industrial/Enterprise IoT
Industrial IoT is expected to capture the 70% of the whole IoT market cap by 2020. In 2016, Industrial IoT enablement gain huge traction and exponential growth are on the cards.
Industrial IoT will fuel the AI adoption. As of now, innovations in AI are tailor made for business/technology vertical. These Things are expensive. Leaders will bet their money on AI which will turn the initiatives to be profitable in the long run. It also enables the leaders to take smarter decisions.
Computing power
Currently, the IoT enablers make smarter decisions (simple rules) in the Things before emitting an event. It will lead to generating smarter data rather than huge amount raw data. Nanotechnology will drive these intelligent decision making in the Things. For mission critical use cases, we need all the data generated from the Things.
Hundreds of billion Things will be enabled in the next 5 years. These Things are going to generate massive amounts of data that we never imagined. We need huge computing power for less cost to process all these data. I believe commercial quantum computing will enable us to tackle this problem around 2020. It is not a theory anymore, as D-Wave quantum computers are real. These advancements will fuel the AI growth exponentially.
Robots
Robots will become common in many places by 2025..If not, at-least in US as per the study. Each Robot is a Thing that is connected to every other thing over the Internet. IoT version 1.0 provides the infrastructure to enable the robots to continue learning like human beings and make better decisions using Artificial intelligence.
Conclusion
Open source projects such as Open AI, TensorFlow are going to accelerate the AI adoption in various ways. When IoT version 1.0 is becoming mature, the AI goes the peak of expectations and hit the trough of the technology cycle. By the time version 1.0 is mature, AI will be ready for adoption in every Thing. I hope it will change the world for better.
Staff ML scientist | ML/DL/AI Engineer |
8y1) In IoT 2.0, I feel Artificial intelligence will move closer to the end devices in the coming years. Rather than raw data, the end devices(Things) may be enabled to reject, categorize, flag the data before sending an event. 2) Regarding the computing power, I wish Quantum Computers with it's extraordinary potential will be commercialized in the next 5-10 years and will favor continuous learning from the huge data generated by the Things. Apart from the points you've mentioned, I feel that the role of AI in IoT v2.0 will be in Security too. A Thing needs to continuously learn to safeguard itself from threats. I think there won't be much problems for security in communication, Internet, back-end as the security in these fields are evolving continuously
Senior Product Engineer at Cognizant
8yI still feel the Domains like BFSI, Healthcare and Manlog are not catching up the momentum considering the security and Privacy as a concern. Recently i have been involved in creating and designing the IoT applications for Banking domain operations. When we tried to Pitch in with the clients they were hesitant to even look at the options. To my surprise the 4 parts you mentioned were considered as a risks by them.