Topics In Demand
Notification
New

No notification found.

Blog
TechVoice: Opportunities in AI Short and Medium Term

1309

0

AI as a field in not new but it was hampered with few Lull periods (AI winters) until back-propagation algorithm was invented that paved the way for present deep learning methodologies.  In spite of having solved the larger challenge of learning the network without handcrafting the features, the availability of data set was hindrance for large-scale deployments of AI.

The dawn of internet and followed by the mobile revolution created much needed data for the algorithms. The reducing cost of computing and reducing cost of storage combined with recent development in the algorithms have made AI as a potential candidate to compare with Industrial revolution. In words of Andrew Ng, “AI is new Electricity”. Business wire mentions, “The artificial intelligence market is expected to reach USD 190.61 Billion by 2025 from USD 21.46 Billion in 2018, at a CAGR of 36.62% during the forecast period”.

Recent development in the field of “deep learning” and “convolutional neural networks” have made good advances in the domains of image recognition, self-driving cars , recommendation engines. The companies have shown the impact the AI based system can bring to increase the sales in online retail (Amazon) or entertainment domain (Netflix). According to a paper written by Netflix executives Carlos A. Gomez-Uribe and Neil Hunt, the video streaming service’s AI recommendation system saves the company around $1 billion each year.

With the advent of Reinforcement, learning i.e. an effective optimization technique the program can take decision where the overall return is higher. This method has yielded very good results in the domain of personalization and giving customized ads and news to the online users.  The potential of personalization is in itself a multi-billion dollar market. One of the estimate reports the proximity marketing in offline world is poised to grow to ~$1.5 Bio.

Deepmind made a breakthrough by combining the Deep Learning and Reinforcement learning to give us the glimpse of intelligence in the games of Atari, Go etc. This makes this field all the more potent for disruption with the combination of multiple algorithms and innovative to ways to fuse algorithms together to produced desired results. With the theme of connectivity around, this will enable to have massive real time data and the need -for algorithms to injest  real time data, process it and take decision autonomously will be desired.

No doubts the processing power requirement for running such algorithms is huge and this has re-ignited the chips industry to create the Application Specific Chips for AI. Nvidia is a case in point, it now have largest market share in GPUs. It reported $1.14 billion of revenue from hardware sold for use in data centers, the first time the segment had cleared $1 Billion in sales up from $634 million a year ago. In addition, cloud service providers find this field as potent in terms of providing cloud based processing for various users and are providing the algorithms and libraries through APIs.

While most of the prominent figures in the deep learning area suggest that that deep learning is fundamentally limited (narrow AI) while human intelligence is not (General Purpose AI). Most of the industry experts in AI area have a common approach to understand the basic principles of intelligence in the human brain and try to mimic or take inspiration to develop bio-inspired architectures and algorithms.  The learning in the human brain is all about connection and reconnection between the synapses and neurons, this provides a hint that brain is a parallel data driven model with little or no software to process. This hypothesis has opened up neuromorphic computing as a new field for developing hardware and chips suitable for AI.

One can only comment the field will grow at a frantic pace as the accessibility and understandability of the algorithms becomes ubiquitous.  Thus, in the short term, the use cases using the exiting algorithms will see massive traction in the industry and this will provide many new organizations to be part of this revolutions.

In my view there will be exponential rise in narrow AI applications in short term while in medium term there will be faster pace of innovation in development of general purpose AI and to our quest to have human comparable machine.

  • Amit Borundiya, Head of Technology, NASSCOM Center of Excellence IoT & AI 

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.


© Copyright nasscom. All Rights Reserved.