Home » The Internet of Things is a Network of “Smart” Objects
The Internet of Things is a Network of “Smart” Objects
Researched by Thomas DeMichelePublished - December 19, 2015 Last Updated - May 11, 2016
What is the Internet of Things (IoT)?
The Internet of Things (IoT) is a network of connecting electronic objects which can collect and exchange data. By utilizing the existing internet structure and smart devices, objects can be sensed and/or controlled remotely, and analytic information can be shared to improve the efficiency of just about every walk of life.
The IoT connects smart grids, smart homes, smart cars, smart phones, smart cities, smart devices, smart glasses, and smart mice’s (including biochips in mice and the mouse you use with a computer). Essentially anything that assists people to use “the internet” directly or indirectly is a “thing” that is part of the Internet of Things.
Like the internet, there is no particular controller of the Internet of Things. In a general sense, the IoT is composed of all the connections between things and the “structured” (organized) and unstructured (unorganized) data from those objects.
Why is the IOT Important?
The IoT allows computer scientists, healthcare companies, finance companies, tech companies, food companies, energy companies, etc. to generate insights and analytics from the world’s data.
The Problem with the IOT
The problem is, 80% of data on the IoT is unstructured. Traditional computers can only make sense of a small fraction of the unstructured data. Every day the amount of data grows and compounds the problem. Luckily companies like IBM have been working on something called cognitive computing that can help humans make sense of the unstructured data on the IoT.
FACT: The IoT’s roots go back to the Massachusetts Institute of Technology (MIT). Work on RFID chips at the Auto-ID Center created initial data. In 2003, approximately 6.3 billion people lived on the planet, and 500 million devices were connected to the Internet. The U.S. census has forecast that, in 2020, there will be approximately 7.7 billion people living on the planet. According to Cisco experts, there will about 50 billion objects connected to the Internet at that time.
IoT: The internet of things, all the connections between devices and all the structured and unstructured data.
Machine Learning: Teaching computers to learn through examples and so they can program themselves, rather than relying on human code to predict every possibility. Machine learning is an important part of cognitive computing.
Cognitive Computing: A computer program that can think like a human. Or rather, artificial intelligence that can think learn, and make connections. Cognitive AI can make sense out of massive unstructured data and reformulate it into simple analytics that can be requested quickly by humans.
Watson: A cognitive computer. He has been in the works for years, and professionals in industries like healthcare already use him (similar to how someone would use Google or Wolfram Alpha). His next mission, to make sense of the IoT. NOTE: Google has their own cognitive AI DeepMind. Right now they aren’t focused on the IoT, so we won’t discuss this more now (but you can learn more cognitive AI like DeepMind here).
Unstructured Data? It’s simply data that hasn’t been organized. Most of the information humans create is unstructured. Classic computers need structured data to process, cognitive computers don’t. They can make sense of unstructured data by learning, thinking, and drawing connections the way humans do.
Imagine 50 billion interconnect objects sending packets of assorted real-time data. Now imagine that about 90% of that data has no organized structure. Imagine that it’s up to Jane, the analyst to make sense of all the raw data and use it effectively. Jane is efficient but human. Creating computer cognition would allow Jane to be the second line of defense, not the first. The analyst of the IoT isn’t Jane; it is IBM’s Watson (or similar cognitive AI).
Watson can reason. It’s adaptive, interactive, inquisitive, and understands context, as a human would. In fact, the idea of AI is to build a machine that thinks like a human but has the increased dynamic memory of a computer.
An analogy between the way cognitive computers like Watson work with unstructured data and the way a human brain works: Humans get bombarded with millions of sensory inputs every moment. This input is like the unstructured data of the IoT. Our conscious and/or subconscious grabs less than ten sensory items at a time to process in short term memory. We make connections between existing information and new data from our sensory inputs. Connections make sense of the data and allow us to decide what is important. Our brains are rewired, storing those connections in long-term memory.
Watson’s short term working memory can theoretically contemplate all “sensory inputs” make connections between all data, store it all in memory, and then recall any connected information at any time in a simple to understand way. In regards to the IoT, it can process all the information on the IoT at once.
The trick to Watson being able to make sense of the flood of random data is that the program runs in a manner that imitates human understanding, thinking, and learning. By using a unique type of machine learning Watson is now able in many ways to mimic the way our memory works and communicate information back to us in ways that humans understand. An excellent example of this in power would be Watson’s attempt at Jeopardy in 2011.
Cognitive IoT. How Watson Works With the Internet of Things
Watson is especially important when it comes to the IoT because Watson is one of the only things on earth that can make sense of so much assorted data. When we pair a cognitive computer like Watson with the IoT, we can call the combination a cognitive IoT.
Of course processing power is limited, so for practical purposes, Watson can be broken down into API’s which are focused on different types of learning and for various kinds of applications. For instance:
The Natural Language Processing (NLP) API lets you interact with systems and devices using simple, human language.
The Machine Learning Watson API automates data processing, continuously learns from each interaction with data and ranks the data based on priority.
How Watson learns. Watson captures data and tries to organize it; humans then curate the content helping Watson understand and organize the data, it then organizes it again in a way that will be useful to humans, then Watson and humans work together to make sense of the data, and the process repeats. This whole process is overseen by experts and subject to their input.
The Video and Image Analytics Watson API uses unstructured data, including data from video feeds and image snapshots, to identify scenes and patterns in video data.
The Text Analytics Watson API mines unstructured textual data (such as Twitter feeds, customer feedback on blogs and transcripts from call centers) to find correlations and patterns in these vast amounts of information.
IBM Watson: Smartest Machine ever built. A documentary from 2014; we have over a year of advances under our collective belt.
FACT: Watson has it’s home base at “IBM’s global headquarters for IBM Watson Internet of Things in the HighLight Towers in Munich”, but Watson can be broken up into API’s to help specific types of businesses.
IBM, building the next generation of Watson APPs.
A cognitive computer like Watson can receive a lot of data and make sense of it quickly. However, to process every possible combination of data from 50 billion plus devices instantly, Watson will be likely to require moving beyond the boundaries of classical computing. Cognitive computing uses classic computers because we don’t have working quantum computers yet outside of the experimental NASA / Google D-Wave. With Watson, a lot of the processing is done using workarounds to avoid limitations of our current technology, as Watson learns and technology grows anything is possible.
Creating Brain Systems, Quantum Computing, Quantum mechanics, and Consciousness.
Read Only V. Read / Write. How Do we Learn from Science Fiction and Ensure Hal Opens the Pod Bay Door?
We have brought the technologies in Sci-fi movies from moonshots to reality. But those movies weren’t just examinations of cool tech, 9 times out of 10 they are warnings of misuses of technology. As silly as it sounds we really do need to look at prophetic works like Star Trek, Minority Report, iRobot, 2001 a Space Odyssey, 1984, Terminator, etc. and ask “what lesson can we learn aside talking AI, rockets, and teleportation is cool?”
Right now cognitive computers are the analysts and the IoT is a wonderland of data that can revolutionize business, education, and healthcare. But, with great power comes great responsibility. It’s important to be thinking ten steps ahead addressing questions like:
What are the dangers of letting a cognitive computer be a controller of devices, rather than just a reader of data? How do you safeguard bad choices? Do analog computers come into play?
What are the dangers in letting a smart AI like Watson do things like control the power grid and missile defense systems? Is there someone just waiting for this to happen?
Would an AI lie to us for self preservation? How would you find out?
How do we generally safeguard against, the more likely issue, a human tyrant using the IoT to essentially control the population (think biochips, nanotech, drones, and advanced analytics)? We can’t forget how often history repeats.
Sure it all sounds like Sci-fi and i’m not the first to compare the IoT and cognitive computers to SkyNet. The question isn’t do we move forward, the question is how do we stay ten steps ahead (even with the help of Watson) to ensure our survival in the long-term in a humane way the ensures individual liberty and collective success?
The Internet of things is smart and connected already. Watson already exists. Quantum computing and cognitive computing exists. Let us hope our good ideas and intentions outpace our missteps.
Thomas DeMichele is the content creator behind ObamaCareFacts.com, FactMyth.com, CryptocurrencyFacts.com, and other DogMediaSolutions.com and Massive Dog properties. He also contributes to MakerDAO and other cryptocurrency-based projects. Tom's focus in all...