Artificial Intelligence: Top five trends for engineers and scientists

Jos Martin has identified five key trends that will be key for engineers and scientists as concepts such as AI, deep learning, data analytics, IoT start to interest and make applications that once seemed futuristic come closer to reality

In 2019, Artificial Intelligence, deep learning, data analytics and IoT will enable applications that once seemed futuristic come closer to reality, fuelled by engineers and scientists tasked with researching and developing AI and other technologies.

As these evolve, so too will developers’ tasks and job roles. Engineers and scientists must learn and adapt to new concepts, design workflows and skills.



AI is becoming more and more widespread across industries, signalling a growing need for it to be broadly available, accessible and applicable to engineers and scientists of varied specialisation. It is not just data scientists that are driving AI experiments and adoption of deep learning in industrial applications. Engineers and scientists have a key role to play too.

Over the next 12 months, complexity of larger datasets, embedded applications, and bigger development teams will drive solution providers towards interoperability, greater collaboration, reduced reliance on IT departments, and higher productivity workflows.

1. Engineers and scientists experiment with AI

More engineers and scientists, as well as data scientists, will experiment with and adopt deep learning in 2019, driven by technical curiosity, business imperatives to reap the promise of AI, and automation tools. AI is becoming more accessible beyond data scientists, due to new workflow tools which are simplifying and automating data synthesis, labelling, tuning, and deployment.

New tools also mean the breadth of applications, from image and computer vision to time-series data like audio, signal, and IoT that are common in numerous engineering domains, are broadening.

Example applications range from unmanned aerial vehicles (UAV) using AI for object detection in satellite imagery, to improved pathology diagnosis for early disease detection during cancer screenings.

2. Specialisation of applications and domains

Industrial applications using AI demand specialisation. Smart cities, predictive maintenance, Industry 4.0 and other IoT and AI led applications demand a set of criteria be met as they move from visionary concepts to reality.

For example, low-power, mass-produced and moving systems require form factors, advanced mechatronics design approaches need to integrate mechanical, electrical and other components, and safety critical applications need increased reliability and verifiability.

Often, these specialised applications are developed and managed by decentralised development and service teams rather than being centralised under IT, creating a further challenge. Examples include agricultural equipment using AI for smart spraying and weed detection to overheating detection on aircraft engines.

3. Interoperability being critical to AI

Interoperability, how computers exchange and make use of information, will be critical to assembling a complete AI solution.

At the moment, each deep learning framework usually focuses on a few production platforms and applications. However, effective solutions require assembling pieces from multiple workflows, creating friction and reducing productivity.

ONNX.ai and others are working to address these interoperability challenges. In turn this will enable developers to freely choose the best tool, more easily share their models, and deploy their solutions to a wider set of production platforms.

4. Edge computing for local processing scenarios

Edge computing enables AI applications in scenarios where processing must be local. Edge computing is set to get smarter in 2019, as advances in sensors and low-power computing architectures take it to the next level with high performance, real-time, and increasingly complex AI solutions.

The technology will be vital to realising safe autonomous vehicle use, as such transportation needs to understand the local environment and assess driving options in real-time.

Edge computing could also enable huge cost savings for remote locations with limited or expensive Internet connectivity, such as deep-sea oil platforms.

5. Fostering of greater collaboration

Uptake of machine learning and deep learning in complex systems necessitates a greater number of participants and more collaboration.
The scope and complexity of deep learning projects, increased by data collection, synthesis and labelling mean larger and decentralised teams are needed.

Systems and embedded engineers require flexibility to deploy inference models to data centres, cloud platforms, and embedded architectures such as FPGAs, ASICs, and microcontrollers. Such personnel also need to be competent with and have expertise in optimisation, power management and component reuse.

Those at the centre of collaboration, engineers developing deep learning models, need tools to experiment and manage the ever-growing volumes of training data and lifecycle management of the inference models they handoff to system engineers.

Jos Martin is Senior Engineering Manager at MathWorks.