AI Outside the Lab: An Overview of Practical Applications in Industry
Advances in artificial intelligence can be broken down into two categories: research and application. These are mutually dependent — research has no value without practical applications, and applications stagnate and become outdated without new research. AI research in Canada is concentrated in 3 universities. The University of Toronto is recognized for deep learning research, the University of Alberta is recognized for reinforcement learning research, while the University of Montreal is recognized for advances within statistical machine learning. Although new advances are intriguing, they are of little relevance to those outside academia. Thus, the focus of this article will be applications of existing AI research.
Traditionally founded by opportunistic academics, startups are the primary vehicles for discovering and implementing novel applications of research with the field of AI. This article examines 3 of the most innovative AI startups in Canada, their specialization within the applied AI space, and how they hope to transform entire industries.
Deeplearni.ng offers a more technical approach to management consulting. Its teams of software engineers, researchers, and consultants designs and deploys deep learning models for its clients, helping them solve traditionally difficult business problems or streamline processes (i.e. automatic recognition of cheques). Supplementing technical expertise with business acumen, deeplearni.ng helps its clients unlock business value by providing them with the most updated knowledge from AI research and tools developed by the academic side of machine learning. For example, deeplearni.ng is working with Scotiabank to help embed deep learning into their credit card and payment collection systems, addressing age-old problems with novel methods based on new research in AI. In addition to its revenue-generating work, deeplearni.ng contributes to the AI ecosystem by hosting meetups focused on educating attendees on optimized implementation of machine learning algorithms. Deeplearni.ng hopes to change how companies think about the value of artificial intelligence by working with them to translate advances in AI research into actionable ways of adding value.
Cortica developed the Image2Text technology to provide a novel approach to digital advertising. Its proprietary software automatically identifies “concepts” in images and video by mapping extracted sentiment to key words. This allows images and videos to be associated with relevant ads, providing digital advertisers with the ability to reach very specific audiences on the web. Cortica hopes to change how advertisers think about interactions with consumers and pioneer a transformative approach to targeting specific customer segments based on sentiment of media rather than categorization.
Sentient Technologies is a startup on the verge of disrupting multiple industries. Sentient Investment Management uses a novel approach to algorithmic trading involving evolutionary algorithms based on human evolution. Sentient Aware personalizes shopping experiences by using recommender systems based on images to analyze buyer intent, getting shoppers to the products they love. Sentient Ascent uses evolutionary intelligence to provide a new approach to testing that improves upon the original AB testing method. Sentient Technologies hopes to show how AI can have a meaningful impact on niche industries traditionally thought to have already discovered best practices.
Overall, the Canadian AI landscape is rapidly expanding and evolving as advances in research and technology increase the feasibility of innovative ideas. These startups represent a fraction of the innovations in AI, but provide a glimpse of the power and impact complex AI research can have. The field is still in its infant stage — there are many more discoveries in AI research left unapplied and so many stagnant industries left not disrupted. Learn more about AI as QTMA continues to publish.
