In recent years, there has been a surge of apps based on machine learning and artificial intelligence. These have been appearing across a whole range of sectors, including healthcare and manufacturing, as well as in the technology sector itself.
What’s more, these technologies are continuing to develop, so what are the AI and ML trends that we should be looking out for as 2019 progresses?
Moving to the edge
Increasingly, we are relying on IoT devices and AI is being used to process the data that these generate. In order to improve response times and provide faster actionable information though, this processing is increasingly being carried out in an edge computing environment.
This allows AI at the edge to deal with unstructured data like speech and video. Examples include analysing security camera footage or providing real-time translations. This type of edge model is also useful in industrial situations, providing for predictive maintenance of machinery, for example.
AI chips
AI needs a lot of processing power, perhaps more than a conventional CPU can provide, and tech companies like AMD, ARM and Nvidia are developing specialist chips in order to speed up the processing of AI tasks. These can be optimised for specific tasks, so you would have different processors for video and for voice, for example.
There are other chip developments too, including Application Specific Integrated Circuits (ASIC) and Field Programmable Gate Arrays (FPGA). These will help in the handling of AI data in order to speed up the creation of databases and allow the data to be queried quickly.
Neural networks
Neural networks – made up of nodes that can transmit information to each other as well as to a central server, thus mimicking animal brains – are likely to be an increasingly key part of AI adoption.
However, to date, neural networks have tended to be self-contained. To exploit AI effectively, it’s going to be vital for neural networks to be able to interact with each other. A number of major cloud providers, including Amazon and Microsoft, have launched the Open Neural Network Exchange (ONNX) to help make it possible to use neural network models across different platforms and architecture. This is likely to become an industry standard for AI systems to exchange data.
Automation
Machine learning and AI are all about automation, but normally it’s necessary to train the system to recognise patterns in order to get things started. A key part of speeding up the deployment of systems will be to automate the ML process itself.
This means that the technology can handle the workflow and process involved, while people in analytics jobs are able to spend more time focussing on the business issues involved.
AI and operations
Finally, AI applied to system operations will assist teams in understanding and managing infrastructure. Applying AI and ML here will allow operations to switch from reacting to situations to predicting where failures are likely to occur. This should also bring together AI and DevOps to boost delivery times and reliability.
© 2022 3ecruit all rights reserved | Privacy Policy | Privacy & Cookie Policy | Powered with 🤍 by Shazamme