On the Intersection of AI and Augmented Reality

While recent attention has been planted firmly in the realms of AI, other emerging technologies have been developing quietly in the background. After the faltering introduction of Google Glass some years ago, the idea of an immediate upsurge in augmented reality applications came to a standstill. While considerable experimentation and new technologies continue to be explored, these initial responses from the consumer population have dampened enthusiasm–at least on the part of pundits and journalists. But exploration of augmented reality continues. Now, with the explosion of AI, and particularly deep learning, it is very likely that augmented reality, virtual reality, and Head-up Displays (HUDs) will become increasingly important.

In previous postings on this blog, we have seen the advance of robotics, and the usefulness of machine learning in this area. We have also looked at hybrid human-robotic systems as well as robot interactions in a human environment. We have posited that the development of human enhancement will continue to parallel the development of smarter artificial intelligence in a kind of continuing evolutionary arms race. Humans will not in the archaic science-fiction sense become machines or absorbed in machines; they will simply be more integrated with easier to use and increasingly powerful AI tools that enhance capabilities and enable greater and more flexible operations in the workplace as well as at home.

In the immediate sense, application of AI to enhanced virtual reality means the possibility of providing things such as smart object recognition on a continual basis with a HUD; for example, identity detection for security, as well as for ordinary human interactions. This will increase the range of identifiable people and enrich conversations by providing details on an individual which might otherwise be forgotten or may not have been known it all. The technology for locating information on individuals is of course widespread on the web; pattern recognition for facial images based on deep learning is already available. Deep learning could also be used to recognize scenarios of various types or components. A view of parts in an automobile, for example, could suggest an immediate repair to a technician, with instructions, tests, and approved tolerances available immediately on a prosthetic device.

Simple pattern recognition enhancements would be only the immediate consequences of AI/VR intersection. AI could be used to guide the hand that performs a function through minute electronic stimulation. A detected image could be used to locate where a problem exists, or where control movement might be required; haptic response could provide the minute instructions needed for immediate learning and reinforcement of a specific action. Similarly, as an aid to handicapped individuals of any level, situational awareness could be enhanced by deep learning associated with cloud-based environmental understanding to produce movement guidance; warnings of balance or functional instability; as well as linking these movements to common physiological metrics such as heart rate, blood pressure, and body temperature.

Application of AI will also enhance all forms of virtual reality by providing recognition of locations and objects in the real world that need to be mapped to the virtual environment. Again, it will be deep learning, associated with external databases of real-world locations, objects, and devices, that will provide the solution.

The capabilities that can be enhanced will greatly improve understanding. By providing such possibilities as extreme long distance sensing in which the capabilities of object recognition and distant camera visioning can provide an understanding of objects at extreme locations for military, traffic and logistics systems with sensory capabilities for prediction of obstacles, threats, and impacts upon efficiency of operation. Information on any remote, difficult, or otherwise invisible location could also provide new sources of data through enhanced object recognition.

The capabilities provided by the combination of AI and virtual reality make it possible to provide a continuous situational awareness that relieves the need to continuously monitor situations within the environment. This will lead to new freedom of movement and will have inevitable consequences in ability to function in new situations as well as to understand the complex nuances and interactions that exist, and yield opportunities within the everyday environment.


Deep Learning in Finance: The Video

Changes are afoot in the Financial Services Industry. New tools from AI and particularly, Deep Learning, are poised to create rapid change across traditional activities. Financial Services have always been highly conservative in IT use, and with good reason. Because of its predictive possibilities, aimed directly at Risk, Deep Learning may open the floodgates to a revolution in this sector.

In this series of videos, you will see a handful of presentations on application of Deep Learning to Finance. As usual, these are all provided on the standard YouTube license, and explanations are lifted from the landing page, or from company web sites where information is not available.

Research to Products: Machine & Human Intelligence in Finance (RE•WORK)

Published on Nov 9, 2016
Talk by Peter Sarlin, Hanken School of Economics, Re-Work Deep Learning in Finance Summit 2016

Artificial intelligence and deep learning in finance has gained traction in the past years. This talk will cover our work in the field of machine learning applied to distress events, networks and news. We look into machine learning for systemic risk identification and distress signalling by measuring excessive increases in micro and macro-financial imbalances, network analytics to account for the interconnectedness of financial markets and deep learning textual data for event extraction with a focus on bank distress in the news.

Deep Learning in Trading (RE•WORK)

Published on Nov 9, 2016

Talk by Hitoshi Harada, CTO at Alpaca, Re-Work Deep Learning in Finance Summit 2016

There are many potential applications of deep learning in financial services, Alpaca has addressed some of them and has been working specifically on the deep learning in trading. In this talk, we will talk about our technique and insights found in our experiments and productions.

Applying Deep Learning at Scale for Financial Applications (CME Group)

Published on Aug 8, 2016
Talk by Amir Husain, CME Group’s Tech Talk 5.0, London

Amir Husain is the CEO and founder of SparkCognition, a company that aims to be at the forefront of the “AI 3.0” revolution.

Deep Learning for Financial Sentiment Analysis (KDD2016 video)

Published on Nov 10, 2016

Talk by Sahar Sohangir, Florida Atlantic University, from the KDD2016 Conference