Statistical Machine Learning versus Rules Based Approaches

Here at Chatterbox Labs we've been at the forefront of championing the numerous business benefits statistical machine learning offers versus a rules based approach.

Gone are the days when you can spend weeks and months building prototype products that are being considered for quarterly releases to address business scenarios that are failing, and expensive to maintain and support.

With the new wave of technology adoption such as IBM Watson, Microsoft Azure, SAP Hana and numerous Marketing, Data and Commerce Clouds, it is evident that every enterprise business is looking to either access or deploy technology that runs in real-time, operates at scale and delivers business outcomes and ROI in timeframes like never before.

In our experience there has always been merit in both statistical and rules based approaches to machine learning. However, the recent shift in addressing dynamic market changes (i.e. to build multi-lingual products within days across languages and industries to disrupt global markets) has truly changed the game and severely impacted the competitive landscape for companies who rely upon rules based machine learning approaches.

The machine learning market is home to a small number of highly skilled companies that have recruited from the best academic institutions with an appetite for disruption on a scale never seen before. Transitioning from a rules based approach towards statistical machine learning requires different skill sets and academic credentials; those companies that have built out large data science teams, frameworks and products with a rules based approach have often taken huge VC funding or been part of a long academic journey before commercial spin out. Sadly, for these companies whilst they have great Data Scientists the market has accelerated to such a point that the methods being applied no longer meet the demand for the real-time world whereby we expect outcomes so much faster.

Why Statistical Machine Learning will prevail alongside Deep Learning & Neural Networks

Statistical Methods

  • Apply to real-time business outcomes
  • Scale and processes faster than rules
  • Minimise manual training & intervention
  • Are inherently multi-lingual
  • Product release & time 2 market is up to 80% quicker than rules

Chatterbox Labs & Market Differentiation

  • 6+ years within academia (Stanford & QMUL) before spin out
  • 7+ years solely focused on statistical machine learning methods
  • Strong Data Science team with Computational Linguistics backgrounds
  • NLP & Machine Learning Framework core to all that we do
  • Path 2 Purchase product suite (patent pending) market ready
  • Strategic partners are building multi-lingual products across industries in days
  • Total addressable market opportunity stands at $116BN as of Oct 2015

Chatterbox Labs is proud to be working with some of the largest and most innovative companies in the world who see exponential value in our statistical machine learning methods.

Plugging the gaps within partner technology roadmaps is what we enjoy doing. The vast differences and business advantages in what statistical machine learning offers versus a rules based approach are clear for all to see.

Unlike 99% of the opportunity that exists today within the technology market, statistical machine learning takes years of experimentation and iteration before being ready for commercial exploitation; even the largest companies in the world cannot replicate and build machine learning companies overnight with the huge resources and monies at their disposal.

It’s humbling and inspirational to know that a journey that started many years ago in academia will be recognised and rewarded for its visionary approach to solving business outcomes quicker, faster and cheaper.

Chatterbox Labs has an unrivaled vision and roadmap for market disruption...

For more information please contact Andrew Watson, VP Strategic Alliances, or visit our web site