Alec Saunders and Mohamed Musbah took the stage at the Canada Fintech Forum to discuss the reality of AI in today's market. Alec Saunders kicked off the session by providing context on the current state of the market and the major technological breakthroughs we have seen in the last decade. Speaking particularly to Natural Language Processing (bread and butter for Maluuba whose focus is on NLP and general AI), Saunders stated that the current error rate with respect to NLP is 5.9% which is about the same as a transcriber. This speaks volumes as to how advanced we've already become in the space.
Saunders stated that we are at a point in history where there is massive change happening across the world, with an estimated $1.2T estimated additional / shifting revenue to be driven by AI in the next 3 years. Relating back to the general buzz of the conference, if organizations truly want to be in the AI realm 10 years from now, they need to be investing in it now.
Delving more into Microsoft's AI strategy, Saunders broke it down into 3 core areas:
- Platform - incorporate AI in tools that are used across organizations (i.e., Azure Machine Learning) to equip them with tools to develop AI in house, and enable deep learning and cognitive solutions
- Infusing AI into existing products being built. For example, MS PowerPoint now offers an option to incorporate subtitles on slides in different languages depending on the language the audience is comfortable with and MS Bing has incorporated AI in their later releases to use social network and query data to predict super bowl outcomes
- Business Solutions - Incorporate AI solutions into accelerators and incorporate them into multiple verticals across banking, health, Telco, Retail, etc. such that new products can be built from the grown up using AI
Musbah took the stage to describe how much Maluuba has grown. He asserted that Maluuba's AI lab is the worlds largest deep learning research facilities for natural language understanding. Maluuba's vision is to solve for general AI.vision to solve artifial eneral intelligence. Important for this crowd, Musbah asserted that AI can be used for Banks and Financial Institutions to personalize their customer experience to keep pace with the market, especially as customer preferences began to evolve rapidly with the emergence of Fintech and exponential technologies.
Musbah proceeded to break down the existing reality of AI in the market:
- We are training pre-programmed models for single, narrow domains
- There is a requirement for a huge influx of training data
- There is a lack of fundamental reasoning or transfer learning capabilities
Do you agree with the MS and Maluuba statements? Do you have a different perspective? Please comment on this post or message me via direct message and we can chat.