A digital twin is an imaginary model that accurately reflects the physical object. These are basically sensors that generate physical performance features such as power output, temperature, and so on. In areas such as manufacturing, digital twins are emerging as a game changer.
To learn more about digital twins and data analytics in general, Analysis India Magazine contacted Vinay Jammu, Vice President of Digital Digital-Digital Technologies.
Register for AWS Data & Analytics Conclave>>
Adapted excerpts from the interview –
Purpose – Tell us about your career. How did you develop the desire to create a career in AI and analysis?
Vinay Jammu: Ever since I saw a program on Artificial Intelligence (II) at National School in the late 1980s, I have been fascinated by how computers capture and represent intelligence. One-legged robots that can be self-balancing using insect repellents and micro-robots that act like insects and move away from humans and open my eyes to the world of entertainment. Here are my professional travels and my top 3 lessons.
A.D. After completing my PhD in 1996, I joined a mechanical technology company and continued working with the IIA to bring industrial prognosis to industrial equipment health prognosis. I have had the opportunity to work with institutions such as Rice University, the U.S. Department of Energy, the U.S. Army Space and Strategic Defense Command, and the US Department of Defense.
A.D. In 1997, I moved to the GE Corporate Research and Development Center and continued to develop advanced AI-based solutions to predict failures and critical equipment such as gas turbines, jet engines, MRI machines, wind turbines and locomotives. A.D. In 2002, I moved from the United States to India, and dozens of researchers were encouraged to lead Prognostics Laboratories in Gee research by developing advanced technology for machine health forecasting.
In 2020, I moved to Gee Digital as Vice President of Physical + Digital Technology and continued to develop domain knowledge + AI-based solutions for GI and its clients.
Purpose – What are your responsibilities in your current role?
Vinay Jammu: For GE Digital for Physical + Digital Technologies In my current role as VPA, my responsibility is to bring Lynn principles and digital technologies together to influence GE businesses and its customers. My team and I are responsible for building wind turbines, image-based damage algorithms for gas turbines, and eye-based data quality improvement algorithms — all of which ultimately help our customers drive better business results and performance. Excellence.
Transforming an organization digitally requires a strategic approach to identifying the right problem solving strategies for the business and getting the right information and AI models to address those issues. Today, most companies do not vote for the company, but for the value of the business. This makes digital transition difficult to measure and maintain.
Crucial in my current role is how we use LIN principles such as Husseini Canary, Value Stream Maps (VMS) and Standard Work to identify change business issues and link them to data and model value maps for systematic and sustainable business transition.
Vinay Jammu: Digital twins are life models of assets, processes, systems or networks. The concept was first developed by Professor Michael Grives and was first implemented in NASA to improve product design and engineering. NASA develops advanced technologies such as spacecraft, Mars rover and many more. Challenges It is difficult to predict the operation of a technology such as a Mars rotator during design. Billions of dollars will be spent in the laboratory to estimate the conditions and procedures for the elimination of as many cases as possible. However, unforeseen circumstances may arise while on a mission.
The purpose of digital twins is to learn from new situations as they occur, to respond as much as possible in real-time, or to develop designs that can better cope with these situations in the future. The ultimate goal of digital twins is to continuously learn new information, so their predictions are as accurate as possible.
Digital twins bring together information-based knowledge and domain-based knowledge using artificial intelligence to build models that produce more accurate results than ever before. These models can be used to solve the most important industrial problems we face today. Digital twins can be used to improve forecasting power from wind turbines to use more renewable energy and move towards net zero carbon emissions. We can build patient twins to drive personalized drugs to improve treatment for diseases such as cancer. As technology advances, the uses of digital twins are unlimited and we will improve the science behind them to solve some of the world’s most difficult challenges.
Purpose – How can digital twins help connect the physical and digital world? What is expected of this technology in the near future?
Vinay Jammu: Digital twins follow the principle of “trash in the trash”. Without accurate information, digital twins will not work well. To do this, digital twins need to interact with the physical world to gather information.
To sum up, we need to work together physically and digitally to address the challenges we face today. In the future, we expect more data and more calculations. Sensors become cheaper, and many of them are incorporated into different asset types, which enhances their ability to produce better information and be smarter. We are seeing a lot of automation happening now, and it is expected to continue. As this improves, we expect that “automated systems” will be transformed into “autonomous systems”, with machines automatically handling tasks that increase complexity and distrust (e.g., self-driving cars).
Vinay Jammu: My PhD work involves developing diagnostic systems to predict and prevent crashes in helicopter gearboxes. NASA has a US military unit at the Cleveland Glen Research Center, designing new helicopter gearboxes. Failure of a gearbox can result in loss of life and significant financial loss. Traditionally, this team has spent millions of dollars trying to predict and prevent these failures and generate possible crash data to improve the safety and reliability of helicopters.
My research was to understand how human marshmallow design uses domain knowledge and how to use experimental data to conduct diagnostic tests and develop human-like systems. In addition, as the helicopter begins to improve performance, the system will constantly learn to use the new information that comes with it. From a simple physics model, I built a “neuro-blurred” signal system that incorporated domain knowledge into the reference system as “blurred” laws. When new flight information arrives, it is constantly being studied using this neural network-based learning algorithm to improve diagnostics. This was in the form of digital twins (although we don’t call it that), and NASA recognized this work by two new technologists. Rewards.
Objective: Which machine learning and artificial intelligence technologies will really hinder the industry in the coming years?
Vinay JammuToday we are at ANI-Artificial Narrow Intelligence, where AI can solve some of the narrowest defined functions, such as facial recognition. The next IE. This is still considered a difficult test with very complex and technical challenges. Finally, ASI is an artificial intelligence that AI can solve problems that humans cannot solve today. Over the next few decades, IE is expected to improve along these lines.
There are a few key areas on these lines that accelerate AI. These include:
- Identify usage issues / problems that need to be addressed for the community
- Data availability and learning algorithms
- Understanding and representing context (including meaning, purpose, right / wrong)
Some of the biggest problems facing our society today – the transition to energy as we move into a net carbon zero economy; Proper health care that provides affordable, accessible and high-quality health care to all 7+ billion people on earth; Security and safety, including cyber security; And efficient transportation of autonomous vehicles, next-generation transportation, and optimized supply chains (including food waste reduction). AI is an important technology that can help solve these problems with the right physical + digital solutions.
Purpose – What is your advice for people interested in ML engineers? What can you do to stand out from the crowd?
Vinay Jammu: I suggest that interested ML engineers first have a solid foundation in science before moving on to advanced AI and machine learning.
Second, there are many good online machine learning courses that one can take to gain knowledge, but it can only be achieved by taking multiple courses or master’s degrees and applying knowledge to real world problems.
Third, there are a large number of free online databases on websites such as Kaggle.com to help data scientists grow themselves. By practicing in these databases, you will learn the importance of domain knowledge in information science.
Lastly, it is important to find a mentor to guide them – the key to making sure they are on the right track.
Join our Discorder server. Be part of a participatory online community. Join here.
Subscribe to our newsletter
Get the latest updates and related offers by sharing your email.