Trends in Artificial Intelligence to Consider for 2018

In 2017, artificial intelligence was all of the rage because of the potential advantages it held for the population in most fields. This is the year that humanity may experience the first tangible effects of the concept. New advances in machine learning and AI courtesy of tech corporations and academic research will be felt in business models. Some of the artificial intelligence trends to look out for in 2018 include the following.

Capsules Artificial Intelligence

Capsule Networks is a type of neural networks advanced by Geoffrey Hinton, a researcher from Google. the method allows for remedying the issues brought by convolutional neural networks which are currently the main standard in image recognition. The CNN approach is serviceable when the images being given are the same as the ones in training.

Source: AI Blog

However, if the same approach is used for images which have rotation or misplaced elements then it would not be able to perform the same way. The capsules approach allows for spatial relationships between the graphical elements and the natural geometric patterns which would be grasped intuitively.

Integration of Voice-based Assistants

During 2017, voice assistants made their way into our homes and some even integrated with the car industry. Nissan, for example, developed a system that utilized Alexa in the home to activate protocols on its vehicles by the user while inside the home. In 2018, a big percentage of the personal assistant users are going to have regular access to different assistants on a simultaneous basis across a number of platforms which means several of the companies will merge to create an all in one user experience.

Source: TechCrunch

This year, vendors will also start to allow people to customize trigger words for the voice assistants thus making them unique to their lifestyles.

Facial Recognition Enters the Finance Industry

Previously facial recognition on a wide scale was impractical considering the expense, lack of adequate technology and the lack of a large database. However, due to the integration of databases, the advance of technology and so the decreasing cost, facial recognition is increasingly becoming used in institutions.

Especially since companies in the retail and tech fields are merging. The use of artificial intelligence means like facial recognition will not be within law enforcement alone but as a means for identity proof during payment of entry to institutions.

Decentralized AI

This would gain traction with edge computing that moves computations from the cloud applications to the edges where certain devices which collect the relevant information are installed.

Source: New Atlas

Firms are thus increasing production of AI co-processors as well as edge neural networks. these would be significant in drone obstacle navigation. this year, there is potential for advancements in low power computer vision, as well as, image signaling hardware engineered to designed to allow AI on edge devices.

Technology will Start to become Empathetic

Practical advancement could allow for a contextual approach to computing with some devices. As such, going beyond the usually discrete responses which currently marks human interaction with machines, there is a possibility of more human-like responses to requests and queries. Conversations would then yield more comprehensive and insightful results.

Source: Robohub

This is a double-edged sword of course because it would be the start of machine sentience which would make them a viable threat being more efficient than the human race and increasingly more powerful.