<html>
<h2>Introduction</h2>
<p>We are all aware that the human brain is one of the most complex organs in the human body, which, among others, has the ability to store up Petabytes of information. This is also apart from the fact that the brain is the seat of cognitive abilities, and also processes memory through interconnection of neurons (very complex; you may say). But I wouldn’t have you ignorant of the fact that artificial intelligence (AI) is advancing with some mind-blowing complexities. You won’t be wrong to infer that that world is about to witness a paradigm shift from the artificial intelligence as usual to “<strong>Artificial superintelligence</strong>” (which means, machines would have a similar cognitive ability with human; even better). In this study, we would be looking at the recent advances that would ultimately herald the era “<strong>Surpassing human intelligence</strong>”. Join me as we take this ride.</p>
<p><img src="https://i.imgur.com/lGf8CUU.jpg"/></p>
<p>[Image Source: <a href="https://pixabay.com/en/artificial-intelligence-ai-robot-2228610/">Pixabay</a>]</p>
<h2>Simulating the human brain</h2>
<p>It has been an age-long desire of scientists to simulate the abilities of humans; which brain-power is one of them. Close enough, we have been seeing AI, but like you may have observed; before now, AI has been majorly programmed with the abilities of “machine learning” which uses a pre-programmed data to relate to experiences in real-time. This works with some forms of algorithms, though the algorithms are not programmed implicitly to follow a particular task (in a layman’s term: it has the ability to learn).</p>
<p>But like we have seen, the learning process is still subject to the existing pre-programmed data which actually works with the neural networks. This is exactly what moved researchers to seek for a way to create machines that would form <strong>independent memory</strong> autonomously (just like humans do), without reference to any existing data. Would you call this “thinking AI?” We will find out soon.</p>
<p>A particular research was heralded by a team of researchers from the University of Bordeaux, Evry Research Centre, and Natural Centre for Scientific Research, which led to the development of an <strong>artificial synapse</strong> (also called <strong>Electronic synapse</strong>) which can learn the same way the human brain does.</p>
<p>We are already aware that <code>the process of learning in the human brain is directly related to our synapses. These synapses work as links between neurons in the brain. So as more synapses are simulated, the links become more reinforced, and the process of learning is enhanced.</code></p>
<p>This is exactly the idea behind the electronic synapse. This is an advancement in “biomimetics”, in which electronics are designed to mimic the functionality of the human system.</p>
<p><img src="https://i.imgur.com/x7rBP6a.jpg"/></p>
<p>[Image Source: <a href="https://www.shutterstock.com/image-illustration/3d-rendering-robot-learning-solving-problems-704499727">Shutterstock</a>]</p>
<h2>Dynamics of Electronic synapse</h2>
<p>The electronic synapse is also called <a href="http://whatis.techtarget.com/definition/memristor">memristor</a>. </p>
<p><code>“Memristor” actually means “memory resistor”.</code> Though this is way advanced than the earlier memristor we’ve known. This particular electronic synapse incorporates an ultrathin layer of ferroelectric chip, which is embedded inside two electrodes (the resistance of this can be triggered using varying electric voltage; in a similar way to the human brain neurons).</p>
<p>This means; in a low resistance, the synaptic link would be increased; and vice versa. This also means that the capacity of the memristor (electric synapse) to learn is actually dependent on the varying resistance in the ferroelectric layers. Also, because of the ultra-thin nature of this layer, it is also referred to as “nanocomponent” chip. Now, the cognitive abilities of the biological brain have been simulated artificially. But what is the future in this?</p>
<p><img src="https://i.imgur.com/kAfJFJv.jpg"/></p>
<p>[Image Source: <a href="https://www.pexels.com/photo/ai-artificial-intelligence-machine-learning-616020/">Pexels</a>]</p>
<h2>Better-than-human AI?</h2>
<p>This is a question that could be brewing in the hearts of many: “<code><strong>would AI out-smart humans in the future?</strong></code>” actually, I would sight some examples from real-life scenarios and leave you to figure out the answer yourself.</p>
<p>A particular research was done at AIVA technologies, which led to the designing of an <a href="https://www.siliconluxembourg.lu/aiva-the-artificial-intelligence-composing-classical-music/">Artificial Intelligence Virtual Artist</a> (Aiva) which was seen to compose classical songs independently of human assistance. Before now, “classical songs” have been seen as a kind of emotional art that is exclusively preserved for humans to compose. But all that idea transcended into oblivion with the creation of Aiva.</p>
<p>If AI can compose songs; which require high level of memory abilities, what can AI not do? Did I tell you that Aiva has launched a musical album titled “Genesis”? Yes! This AI has a musical album, with some other singles too. Are you now getting what I mean when I normally that “AI could out-smart humans”? This is not to mention some other places where we have seen them beat the human abilities. I believe the world of AI is moving from just “machine learning” as usual to “deep learning”, which would ultimately lead to “artificial superintelligence”.</p>
<p>Can I tell you this? <code>A point is coming when AI would be able to develop independently without any intervention from the humans. </code>A prediction was made by HP’s chief technology officer; Shane Wall, who is also a renowned futurist:</p>
<blockquote>AI will out-smart humans by the year 2029. – Shane Wall. <a href="http://www.business-standard.com/article/current-affairs/robots-will-outsmart-humans-by-2029-11-years-earlier-than-expected-hp-cto-117091300177_1.html">Source</a>.</blockquote>
<p>This is actually some years earlier than the proposed singularity by the futurist; Peter Diamandis (co-founder of Singularity University). But would these predictions hold? Judging by the trend of technological advancements, you can’t say the least. But in my opinion, the waves have already started, and these may arrive even sooner that we think.</p>
<p><img src="https://i.imgur.com/RaKjXNe.jpg"/></p>
<p>[Image Source: <a href="https://pixabay.com/en/robot-artificial-intelligence-woman-507811/">Pixabay</a>]</p>
<h2>Conclusion</h2>
<p>“Learning” has been an absolute prerogative of brains to perform, but with the advances in biomimetic technology, AI has been seen to be able to simulate the same functions; thanks to electronic synapse.</p>
<p>I would still leave us with this question which I often ask:</p>
<blockquote>What can AI not do?</blockquote>
<p><br></p>
<p><em><strong>Thanks for reading</strong></em></p>
<blockquote><strong>All images are CC0 licensed and are linked to their sources</strong></blockquote>
<blockquote>References for further reading: <a href="https://www.nature.com/articles/ncomms14736 ">Ref1</a>, <a href="http://www2.cnrs.fr/en/2903.htm">Ref2</a>, <a href="https://www.technologyreview.com/the-download/610038/an-artificial-synapse-could-make-brain-on-a-chip-hardware-a-reality/">Ref3</a>, <a href="http://news.mit.edu/2018/engineers-design-artificial-synapse-brain-on-a-chip-hardware-0122">Ref4</a>, <a href="http://www.zdnet.com/article/artificial-synapse-creation-makes-brain-on-a-chip-tech-closer-to-reality/">Ref5</a></blockquote>
<p><img src="https://i.imgsafe.org/ae/ae9cc89fec.jpeg" width="239" height="164"/></p>
<p><img src="https://steemitimages.com/DQmf18V3PQenbUzYfkZcWSVGgmV5CtZNymi824SYwAQ2GTp/DQmf18V3PQenbUzYfkZcWSVGgmV5CtZNymi824SYwAQ2GTp.gif" width="1000" height="253"/></p>
<p>gif by @foundation</p>
</html>