With the increasing need for operational efficiency and process automation across the board, from small-scale projects to the magnitude of smart cities, the demands for data continue to become more exacting. Steve Harris, CEO of Mindtech, explains how achieving them with real-world data is challenging, thus launching synthetic data into the spotlight to address these shortfalls. The potential for uniting synthetic and real-world data in the pursuit of smart cities is highly promising for the future of smarter AI.Â
AI has proven its worth in driving innovation time and again, but if we mean to deliver on growing demands and address the shortcomings of real-world data, how ML models are trained needs to be re-examined. Annotations for specific use cases are critical to enhancing the implementation of AI. Until recently, this task was largely dependent on humans, making it costly and time-consuming. This, combined with qualms regarding privacy issues and deployment for anomalous situations, highlights the need for a better solution.
Gartner predicts synthetic data will overshadow real-world data in AI models by 2030. Generated by a computer simulation, synthetic data can mimic features of real-world data and statistically and structurally match it. The scarcity of anomalous event data can be addressed through the ability to create any edge case imaginable. Its machine-generated nature saves valuable time and energy while also providing the opportunity to counteract incumbent bias in the base data.Â
Critically, model training should not be limited to one data source or the other. The two can be complementary in efficiently progressing training at a lower cost, with greater applicability and without privacy or bias issues. But, it stands that synthetic data is essential for developing smarter AIs capable of recognizing and preventing corner cases in smart cities.
Where Does Synthetic Data Fit Into Smart Cities?
Smart cities have been a topic of discussion for decades. The concept places citizens’ quality of life at the utmost precedence, aiming to transform for greater efficiency. This is through the use of technology that is meant to bring increased safety and sustainability. However, there seems to be no shortage of street crime, traffic accidents and urban pollution. Evidently, there remain strides to be made.Â
The technology exists to make cities greener and safer. It is largely a matter of deployment, and synthetic data plays a key role in this. If we wish to see modern smart cities, vision systems are in need of precisely annotated, high-quality, and privacy-compliant data in large quantities. As discussed, this is not always readily available, particularly if we rely solely on real-world data. In fact, it is unlikely to be readily available.Â
Synthetic data presents several advantages for training these systems that can go on to make more informed decisions. Constraints in using sensitive data are not an issue, datasets can be tailored to edge cases that may not otherwise be feasible to obtain, and insights can be gained inexpensively and quickly. Biases can be minimized, further enabling improved predictive modeling. It is ultimately an essential tool in driving the adoption of AI for smart cities.
See More: 5 Ways Data Scientists Can Advance Their Careers
Safety Is the Number One Priority
Fundamentally, smart cities are safer cities. What scares people most in these environments? Arguably, it is random acts of violence, gun or knife crime. Also, with the rise of hybrid working, employees can choose to work anywhere they please. This then incentivizes public officials further to ensure investments are made to reflect well in terms of their cities being considerate of citizen wellbeing, from personal safety and public health to extreme events like natural disasters.
While real-world data alone can be enough to deal with typical, easily predictable scenarios, in the case of protection, there is a gap for dealing with anomalous situations. Control centers consist of hundreds of monitors streaming video. Human controllers cannot be responsible for surveilling every screen. An AI system trained to capture and recognize incidents can greatly free up manpower for tasks requiring human supervision. Therefore, data will play an increasingly crucial role in crime prevention.
Implementing smart city technology changes the way operations occur, whether this is through license plate recognition and gunshot detectors or body cameras and crowd monitoring. Data and AI can transform how safety protocol is conducted. Proactively reimagining public safety will go on to build stronger communities and mutual trust between governing bodies and citizens, but this is reliant on technology that can be flexible. AI solutions powered by synthetic data are the key to achieving this by helping to gain a holistic understanding of the population and deepen public confidence in city officials.
Smarter Is Better
Smarter cities are intended to be better to live in — being able to easily find a parking spot, not constantly being stuck in traffic jams, roads being fixed on time, efficient waste removal, and so on. This focus on technology should come with a human focus — not at the expense of people.Â
The capacity to collect larger volumes of information, through synthetic data, across a range of scenarios builds a truly intelligent system prepared for anything. Being able to transform this into actionable insights is essential to smooth running. Synthetic data-powered AI accelerates the journey from information to action, providing insights into understanding the community’s most urgent challenges and the ability to react swiftly.Â
Smart lighting is a common entry point for implementing smart services. Deploying adaptive technologies like energy-efficient streetlights that detect human presence and consume energy only when needed can significantly reduce electricity consumption. Synthetic data can fine-tune efficiency, account for various situations, including weather conditions, and help improve security and reduce electricity bills substantially. This not only streamlines the procedure but also highlights the need for sustainability to be at the forefront of the agenda.
Since synthetic data is not subject to the same privacy issues as real-world data, there is a greater opportunity for transparency. Citizens should be at the heart of service-building, and synthetic data makes this a reality. This ensures optimization but also that actions align with citizens’ needs.Â
Does This Come at a Cost?
At the end of the day, convincing the powers that be to implement technology will come down to how much money it will save long-term. Smart cities come at a cost — a huge one — but they will eventually be more economical. Real-world data is costly not only in terms of monetary expenses but also in resources and time. Synthetic data can be more cost-effective to generate and readily available for any edge case needed.
It is important to acknowledge that while a diversity of datasets is needed if we wish to optimize training models, synthetic data will play the pioneering role in generating smarter AI in the pursuit of globally smarter and safer cities.
Have you used synthetic data for any of your projects? Let us know how it has helped on FacebookOpens a new window , TwitterOpens a new window , and LinkedInOpens a new window .