DEEP LEARNING….....A State Of the Art
Updated: Apr 3
Have you ever thought of someone who will always be there for you, who will listen to your words and understand you ? Or Have you ever thought of giving birth to a child and feed him with the fruit of knowledge?
If your answer is NO then hold your moment, close your eyes and you can tell yourself - IT’S OK TO LIE(No one will listen what you are saying inside your mind). I believe deep down you always know the answer of these question are yes and I can see you are imagining a relaxed life where you are chilling out with someone and someone is working for you.
WHAT IF YOUR IMAGINATION BECAME TRUE?
YES, you are thinking right. It is possible. And credit goes to Geoffrey Everest Hinton who has been advocating for a machine learning approach to artificial intelligence since the early 1980's, looked to how the human brain functions to suggest ways in which machine learning systems might be developed and found the new generation of technology DEEP LEARNING.
“ In A.I., the holy grail was how do you generate internal representations. “
-Geoffrey Everest Hinton
Deep learning is more than just a souped-up regression It is getting a lot of attention these days, and for good reason. It’s achieving unprecedented levels of accuracy—to the point where deep learning algorithms can outperform humans at classifying images and can beat the world’s best GO player.
Deep Learning is a new area of Machine Learning research, which has been introduced with the objective of moving Machine Learning closer to one of its original goals: Artificial Intelligence.Similarly to how we learn from experience, the deep learning algorithm would perform a task repeatedly, each time tweaking it a little to improve the outcome. We refer to ‘deep learning’ because the neural networks have various (deep) layers that enable learning. Just about any problem that requires “thought” to figure out is a problem deep learning can learn to solve.
A successful and popular version of machine learning that uses back prop neural networks with multiple hidden layers. And I have a feeling, 2019 was way intense than years before. The 2012 success of Alex Net, then the best machine learning network for object recognition, was the tipping point. Deep learning is now ubiquitous in the Internet. The idea is to have each layer of processing perform successively more complex computations on the data to give the full multilayer network more expressive power. The drawback is that it is much harder to train multilayer networks.It is not clear, given what we know about neurons and neural plasticity, whether a backprop network can be implemented using biologically plausible circuits. However, there are several promising efforts to implement more biological plausible learning rules, such as spike-timing–dependent plasticity.
The technology is here, we are connected, and ideas exchange is as never before. And the best thing about AI Renaissance is: popularization and democratization of DL/ML. Nowadays not only Python speakers and NVidia GPU owners can enjoy the sheer endless possibilities: everybody can do it. Writers, Artists, people of other non-tech defined fields can use Colab/Jupyter Notebooks, user friendly applications like Artbreeder and RunwayML etc.
By- Priyanka P. Pattnaik
Intern of CoE-AI,CET,BBSR