Fractal Analytics defines 'speed reference' for AI engineering & design
Artificial Intelligence (AI) for human decision-making software company Fractal Analytics this week held its ai.lcy (pronounced: AI-el-see-why) event in London.
The event, which is the third of a series, included a number of sessions and roundtable discussions and was hosted at The Langham Hotel in London this February 2019.
Artificial Intelligence (AI) for human decision-making software company Fractal Analytics this week held its ai.lcy (pronounced: AI-el-see-why) event in London.
The event, which is the third of a series, included a number of sessions and roundtable discussions and was hosted at The Langham Hotel in London this February 2019.
We hear a lot about this term web-scale (as in, ability to scale any [software-based] operation, workflow or function upwards to what could potentially the size and scale of the web itself)… and so Fractal Analytics takes that thought (literally, in the AI sense) forward and talks about ‘problem-solving’, that is – the ability to apply intelligence to business through a combination of AI, engineering and design.
The event’s keynote speech on how AI has revolutionised business strategy was delivered by Kenneth Cukier, Economist senior editor and co-author of ‘Big Data: A Revolution That Transforms How We Live, Work and Think’.
Data pipelines
Other sessions provided hands-on advice on topics such as: where and why to use AI to solve organisational problems created by lack of resources; and how to build data pipelines to enable AI and scale algorithms to power decisioning in the enterprise.
“AI is here. It is real and is evidently transformative, but AI alone is ‘arguably insufficient‘ to tackle enterprise problems at scale,” claimed Sankar Narayanan, chief practice officer, Fractal Analytics.
Examples of where Fractal has built several AI product startups within its ecosystem across industries include:
- Qure.ai – a software system designed for radiologists when making diagnostics decisions. Qure’s qXR solution is the first AI-based chest X-Ray interpretation tool to receive CE certification. Qure’s recently-launched qER solution for interpreting head CT scans has been proven to have more than 95% accuracy in identifying abnormalities.
- Cuddle.ai – a tool to change how executives make tactical and operational decisions. Cuddle automatically alerts users on what they need to know about their business and allows them to ask questions about their business in natural language.
- Trial Run to improve strategic and operational decisions. Trial Run is a cloud-based experimentation platform that enables companies across verticals to test, measure and refine strategies before rollout.
A ‘speed reference’ for AI
CEO and co-founder at Fractal Analytics is Pranay Agrawal. Speaking at the London event itself, Agrawal explained how it’s important to be able to grasp some form of ‘speed reference’ on data and its surrounding world if we are going to be able to truly extract intelligence value from it.
“Working out what is really happening [with data and being able to truly interpret its contextual meaning] often comes down to being able to gain some level of speed reference on data so that you know what it means in the context of the world around it — think about the fact that when you are in a plane, it’s almost impossible to be able to appreciate how fast you are traveling, unless perhaps you happen to pass some very nearby clouds,” said Agrawal.
The upshot of this, for Agrawal and his company, is that AI needs to be a process of not just number crunching and (if you will) some notion of ‘pure AI’ in mathematical or pure algorithmic terms, it also needs to be a process where contextual relevance and value is created through engineering and design.
Agrawal further used his stage time in London to talk about the fact that ‘control’ (of data, or indeed of any entity) does not always lead to optimisation per se… so knowing what level of AI and analytics you are adding to a so-called ‘choice architecture’ is a crucial part of knowing what kind of decision model you are building.
Fractal has more than 1,200 consultants spread across 15 global locations including the United States, UK and India.
So overall then, the company appears to have justified the need for a new confluence between ‘not just AI’ and data… but AI plus engineering plus design as the core prerequisite for real AI intelligence.
