Topics In Demand
Notification
New

No notification found.

AI Innovations in the last decade
AI Innovations in the last decade

July 6, 2021

AI

237

0

We are now living in the Decade of Deep Learning and AI. This is the beginning of one of the most powerful technologies that can immensely change our lives.

With the large amount of data produced by mobile phones , neural networks became really useful in categorizing ,learning and drawing predictions from data. We have witnessed countless applications breaking records. Beginning from autonomous driving, face recognition to smart home devices.

 

The goals of deep learning ,machine learning are to build more complex and human-like models, which is primarily a transition from weak AI to strong AI and implement human intelligence in machines and applications. For this same motive ,numerous algorithms are carried out and now these algorithms function the world we live in.

 

At a glance ,over the past decade we have already accomplished a lot which is seen in the amount of applications we use in our daily lives. The obvious one being the increase in the use of mobile phones and replacing of televisions and radios. Neural network innovations provided cameras with better quality pictures with face recognition software ,social applications with better web searching tools which created larger social groups.

 

One such innovation in the past decade was the voice-enabled assistance. Although there are more barriers that this system has to improve on .Some of them being ability to store conversations for a much longer time and use this to build facilities to anticipate user requests. The user should be able to provide feedbacks and the artificial system learns and adapts. Further, smart homes and autonomous vehicles and robots can not only take inputs but also predict the next move. Thus neural networks ,an indispensable part of AI revolutions ,are algorithms that create artificial frames that transform data to decisions. Creation of neural networks can be a challenging task when defining cost functions ,calculating error between predicted and true labels.

 

A major limitation in the current neural networks is their insufficient predictive power. In order to interact in a complex environment ,building predictive models is as much important. Reinforcement learning helps to do this. It teaches machines to self-learn and increase its predictive power. Thus in the next decade, there will surely be an increase in algorithmically created neural networks and a reduction in the number of human designs of neural network architectures.


That the contents of third-party articles/blogs published here on the website, and the interpretation of all information in the article/blogs such as data, maps, numbers, opinions etc. displayed in the article/blogs and views or the opinions expressed within the content are solely of the author's; and do not reflect the opinions and beliefs of NASSCOM or its affiliates in any manner. NASSCOM does not take any liability w.r.t. content in any manner and will not be liable in any manner whatsoever for any kind of liability arising out of any act, error or omission. The contents of third-party article/blogs published, are provided solely as convenience; and the presence of these articles/blogs should not, under any circumstances, be considered as an endorsement of the contents by NASSCOM in any manner; and if you chose to access these articles/blogs , you do so at your own risk.


Student

© Copyright nasscom. All Rights Reserved.