Now Reading
MATLAB EXPO 2018: How MathWorks Is Making AI Accessible

MATLAB EXPO 2018: How MathWorks Is Making AI Accessible

Abhijeet Katte

Analytics India Magazine attended the Media Roundtable on the subject, “Are You Ready for AI? Is AI Ready for You,” held at the recently-concluded MATLAB EXPO 2018. The panel was headed by Michelle Hirsch, head of MATLAB Product Management at MathWorks, Sunil Motwani, industry director at MathWorks India and Prashant Rao, technical manager at MathWorks India.

MathWorks And AI:

Rao said that the long-term vision of all MathWorks products was to make AI accessible and easy, build products that are simple to integrate with everything else and create easy-to-deploy tools which can be used in both cloud and edge-based computing scenarios. Several successful user stories were also presented.

Hirsch said that one of the main differentiator of MathWorks products in artificial intelligence is that MATLAB fits in with all stages of data-dependent intelligent systems. MATLAB can be used in stages of accessing data analysing data, model development and deployment. The deployment can be done both with embedded scenarios and cloud-based solutions.



Over the course of discussion, Motwani mentioned one of the highlights of MathWorks products was that its users can create and test failure scenarios. Many clients and users have found the power of model building and testing using Simulink to be very useful and have helped them to avoid many edge cases, building much safer systems. He also mentioned that the main focus going forward is in the areas of image processing, automated driving, data analysis and predictive maintenance. Rao added that the analysis of time series data will be an important value add going ahead. When asked about the kind of clients that MathWorks caters to, Motwani mentioned that the clients are broad range of companies ranging from manufacturing, IT services companies and also from the banking sector. Researchers and students are still a popular and growing segment of users who are adopting MathWorks products, he went on to mention.

Use Cases:

Denso Corporation a global automotive components manufacturer is using MathWorks products to minimise home energy costs. Respiri is making use of MathWorks products to eliminate error prone self-reporting in case of asthma patients. Safran, a multinational aircraft engine, rocket engine, aerospace-component and defense company is using MathWorks products to increase aircraft availability, reduce maintenance costs using real time data and analysis.


W3Schools

Jump-Starting Startups:

Rao also spoke briefly about their support to the startup community. He shared the company’s experience of working with and helping Team Indus. The startup has made it its mission to send a robotic explorer to the moon. When asked about competition between many deep learning frameworks like TensorFlow and Caffe and how it affects MATLAB, Hirsch mentioned that they are in a state of co-opetition (an amalgamation of cooperation and competition) with all the frameworks and they are happy to work with other companies actively to make machine learning easier.

See Also
How AI-Powered Crowd Counting Is Bringing Clarity To Hong Kong’s Political Turmoil

MathWorks And Education:

Hirsch then talked about how MathWorks generated $1 billion dollars in revenue in 2017 and around 3 million users globally. The prime products of MathWorks are MATLAB and Simulink. MATLAB is a tool designed for analysing data, developing algorithms and creating models. On the other hand, Simulink is used for designing and simulating your system before moving to hardware. The tool helps to explore and implement designs that you wouldn’t otherwise consider – without having to write C, C++, or HDL codes.

Hirsch went on to say that MathWorks is currently working with over 5,000 universities, 1,000 schools worldwide — of which 87 are in India. Additionally there are around 1,800 books that use MathWorks products as part of them. Hirsch also stressed that the company was working hard to integrate MATLAB with other software engineering components like GitHub, among others. There are already 1,20,000 submissions on GitHub relating to MATLAB.

 

Provide your comments below

comments

Copyright Analytics India Magazine Pvt Ltd

Scroll To Top