Disclaimer: All these points mentioned are from my own personal viewpoints.
I’m guessing you have recently been seeing a lot of AI posts on Facebook or Twitter and wondering why. There is a simple explanation why AI is used as a modern statistical tool to solve problems in almost each and every field like Astronomy, Biomedical, Customer analysis, Finance, Fashion, Market strategies, Games, Research, Technology…. this list continues. You might wonder how can this AI tool be used to solve all of these problems? Yes, my friend, all the problems have multiple approaches to solve, the best approach when you don’t understand the process among all of them is Data-Driven approach. Where you don’t need to know more about the physics of the problem, but rather should have the capability to build the mathematical models which learn important features which define physics on its own. These models fall under a broad category of Artificial Intelligence. Today, we are facing a new wave of radical change on cognitive automation, which could be broadly summed up under the keyword “AI”. And like any time in the past, we are worried that this new set of technologies will harm us — that AI will lead to mass unemployment, or that AI will gain an agency of its own, become superhuman, and choose to destroy us. Today most of the top companies like Google, Facebook, Amazon… are trying to build models that depict human-like behavior in specific tasks, on the other hand, companies like kernel, neural link… are behind, increasing human brain bandwidth with the help of human-machine interfacing technology, they are trying to decentralize the knowledge source each and every human. It sounds scary, isn’t it?
But what if we’re worried about the wrong things, What if the current world needs AI to address many problems to keep continuing. What if the real danger of AI was far away from the state “super-intelligence” and “singularity” that many are panicking about today? According to me, the Ultron lifting hammer needs at least 10-20 decades or not even possible. As being Biomedical Engineering student, I feel AI is necessary to handle the current state of society, with an overloaded population, very new diseases, which for us as humans would be very hard to find correlations or analyze, until and unless you are trained to do so. But AI models can help to find outliers in the populations just without even having a deeper understanding of the problem. Whose results can be further refined by human experts? In this whole process, AI helps in saving tons of amount of time and money which is usually spend in unproductive work as experimentation, simulations and so on. Don’t get me wrong, I’m not saying AI is final decision maker, It can only help us to reduce search space drastically thereby cutting down cost and time.
Why sudden phase shift in industries?
The huge shift in interests among researchers, industries is due to the evolution of new generation parallel computing devices. For example, until 2015, all ad targeting algorithms across the industry were running on mere logistic regression. In fact, that’s still true to a large extent today — only the biggest players have switched to more advanced models. Logistic regression, an algorithm is one of the most basic techniques in machine learning. It is the reason why so many of the ads you see online are desperately irrelevant. Machine learning and AI have been making fast progress in recent years, and that progress is only beginning to get deployed in targeting algorithms and social media bots. The use of Deep learning methods has only started to make its way into chatbots, newsfeeds in 2016. It is quite striking that companies like Google, Facebook and Microsoft have been investing enormous amounts in AI research and development, with the explicit goal of becoming a leader in the field. The effect of more advanced techniques like Reinforcement learning in this field needs lots of exploration, who knows it may even beat current state-of-the-art.
Few use cases of AI…
**Biomedical: **
Very interesting experiments and researches are taking place in the field of biomedical engineering using the tools of deep learning and machine learning. Computational Biology, today is emerging at a very high progress rate, lot’s of research groups, institutions and companies are exploring these data-driven approaches to study many aspects of humans in a non-invasive manner. In my knowledge I have seen people using AI tools, for genomics analysis, Tumor analysis based on MRI or CT, abnormalities in heart rhythms, depression analysis, to treat Parkinson and Alzheimer’s, effective image reconstruction, Visual defects due to DR, and many more.
**Finance: **
Another interesting application of AI is in the field of finance, market prediction and analysis is one of a very hard task. The main difficulty in such kind of problems is the involvement of human factors, which is highly stochastic in nature and makes the problem hard to analyze. Base on current advancements in AI, with the help of bidirectional LSTM, NLP and recurrent reinforcement learning these analyses and prediction can be done to an extent with lesser than 5% of MSE error. Companies like Morgan Stanley, Goldman Sachs, HSBC, and many other companies use these econometric techniques to increase their revenue rate.
There exist tons of other use cases of AI, I’m stated these two because I worked in these areas and have a greater understanding of AI concerning these fields. In all these applications AI serves are primary analyzer still human experts sits on top of AI agent to verify the outputs. By looking at all these applications it’s very correctly said by Andrew Ng that AI is new electricity at least in the current generation.
Conclusion
We should build AI to serve humans, not to manipulate them for profit or political gain. If you think your jobs cannot be done by an AI, do some google research and think again. If we think carefully about AI, One path leads to a place that really scary, and other leaders to a more towards the development of humanity, science, and technology, There’s still time to take the better one. If you work on these technologies, keep this in mind, your choices affect us all. And you may eventually be held responsible for them.
Feel free to add comments and also share your thoughts… Thank you for your time.
Leave a Comment