The evolution from data annotation to digital twins marks a significant shift in manufacturing technology. Initially focused on labelling data for AI training, data annotation has transitioned to constructing comprehensive digital replicas of physical assets or processes, known as digital twins. This transition represents a leap from mere data labelling to creating complex simulations that provide real-time insights and predictive capabilities, revolutionizing the manufacturing landscape.
Radha Basu is a prominent figure in Indian tech entrepreneurship, best known as the Founder and CEO of iMerit and co-founder of Anudip Foundation. With a career spanning leadership roles at SupportSoft and Hewlett Packard, Radha has been dedicated to addressing youth unemployment by providing digital and data services training. Her initiatives have empowered hundreds of marginalized individuals, reflecting a commitment to technology and social impact. Radha also serves on several boards, including NetHope, Jhumki Basu Foundation, and Santa Clara CSTS.
The interview examines the origins and evolution of data annotation services, considering the influence of generative AI on business models and the synergy between technology and human labour. It explores the transformation of these services, especially in manufacturing, highlighting advancements and changes observed over the past decade. It also delves into the significance of digital twins in predictive capabilities, fault detection, and integrating diverse technologies in smart manufacturing.
The manufacturing landscape has undergone a seismic shift propelled by advancements in technology. What began as data annotation for AI training has now evolved into creating digital twins – comprehensive virtual replicas of physical assets or processes. This transition signifies a leap from labelling data to constructing intricate simulations offering real-time insights and predictive capabilities, reshaping industries’ operations.
Radha Basu’s Vision:
At the forefront of this transformation stands Radha Basu, a luminary in Indian tech entrepreneurship. Her roles as the Founder and CEO of iMerit and co-founder of Anudip Foundation reflect a commitment to technological innovation and leveraging technology for social good. Basu’s initiatives have empowered marginalized individuals by providing digital and data services training, demonstrating technology’s potential to foster societal impact.
She said, “We started about five or six years ago, looking at how we can contribute to the veracity and the more stringent way of building models.”
Origins of Data Annotation Services:
In a candid interview with AIM Research, Radha shed light on iMerit’s journey. With the company celebrating its 10th anniversary, Basu reminisced about the initial focus on refining model accuracy and performance. “We started with building models, and we decided that if we set up technology, developing software products and stuff, we set that up and had the best expertise and the techniques of bringing together the technology, the talent, the expertise and the technique.”
The company’s foray into data annotation stemmed from the desire to build AI models that could seamlessly transition into production, stressing the pivotal role of meticulously annotated data in achieving this goal.
Impact of Generative AI:
The advent of Generative AI marked a significant turning point for iMerit. Basu debunked fears of job displacement, instead embracing Generative AI’s potential to streamline and augment data annotation processes. She said, “Generative AI is going to challenge unprepared businesses. Our role with the tech, especially with the Ango Hub, now that we purchased and integrated the tech, is to provide reinforced learning at every stage, with human and expert feedback at the crux of generative AI.”
This technology became instrumental in crafting purpose-built applications, significantly enhancing defect detection capabilities across manufacturing phases by automating repetitive tasks and empowering human annotators to focus on nuanced decision-making.
Human Element in Data Annotation Services:
“In the decision making… where safety is concerned, scaling is concerned with trust and auditing of the model, especially in production. When I say production, I discuss MLOps broadly, monitoring, test case scenarios, and validation.”
Despite technological strides, Basu emphasized the enduring importance of human expertise in data annotation. From basic annotation tasks to intricate decision-making, human involvement remained integral, particularly in safety-critical scenarios, scaling operations, and validating models. Basu underscored the evolution from mere annotating to becoming the driving force behind technological advancements in data processing.
Evolution of Data Annotation Services:
Over the last decade, iMerit’s data annotation services have undergone a profound metamorphosis. Basu elaborated on the transition from rudimentary defect tracking to constructing purpose-built applications tailored for defect detection across various manufacturing phases. The fusion of expert judgment, anomaly detection, and customized workflows substantially reduced defects during production, optimizing operational efficiency.
“We’re not just doing the annotation; we’re generating the test case scenarios… Conceptually, think of it as creating this digital twin of a manufacturing floor.”
Significance of Digital Twins in Manufacturing:
Basu expounded on the pivotal role of digital twins in predictive capabilities and fault detection. These virtual replicas, combining simulations, human expertise, and IoT sensors, are invaluable for analyzing manufacturing flows and improving operational efficiencies. She mentioned, “You are using the simulation to see the effects of these changes you’re making and the defect tracking. It’s defect detection and tracking.”Basu stressed the importance of integrating digital twins into production processes, offering insights into enhancing productivity and reducing errors.
Integration of Technologies in Smart Manufacturing:
Drawing parallels across industries, Basu elucidated the seamless integration of technologies in smart manufacturing. The synergy between AI, IoT, and human-in-the-loop systems, highlighted through real-world examples, showcased how these technologies collaborate to optimize efficiency and safety across diverse sectors. Basu emphasized the need for continual integration and improvement to drive advancements in manufacturing. Radha emphasized, “The synergy between AI, IoT, and human-in-the-loop systems showcased how these technologies collaborate to optimize efficiency and safety across diverse sectors.”
Conclusion:
The evolution from data annotation to digital twins represents a watershed moment in manufacturing technology. Radha Basu’s insights illuminate the transformative power of technology when coupled with human expertise. This fusion fuels innovation and lays the groundwork for a more connected and intelligent manufacturing ecosystem.