"A Computer would deserve to be called intelligent if it could deceive a human into believing that it was human" - Alan Mathison Turning.
by Som Dutta, CTO & Co-founder of Resonance Data (Rene’)
Origins of NLP
Natural Language Processing has come of age since the days of Alan Turing. However, its origins can be traced back to the 60’s, when the focus in the early years of the cold war was to build systems that would translate Russian into English for snooping American spy agencies. Since that failing effort, there were some promising developments in the area of conversational agents.
Early Days of NLP
ELIZA was the first chatbot, to mimic a psychologist in its interactions with users. ELIZA was reasonably convincing, but like any NLP system built up to the late 80s, it was essentially a patchwork of rules and hand-tweaked parameters. It was only when statistical NLP and machine learning techniques started to take off in the 90s, that NLP began to come into its own.
Where is NLP today?
Over the years NLP has found its applications across a wide array of use cases.
ImageNet Moment In NLP
All these areas have benefited from the steady advances in deep learning over the last decade. Some technologies like Neural Machine Translation have made massive leaps in quality and effectiveness of NLP engines. However, everyone has been waiting for the ‘ImageNet moment’ in NLP. ImageNet is an annual competition that turned out to be a breakout moment for computer vision in 2012. The 2012 ImageNet competition was won for the first time by a Convolutional Neural Network (CNN) that used deep learning to beat all existing benchmarks by a wide margin. Most of the advances in computer vision since then owe their success to the CNN architecture that won ImageNet in 2012.
A similar breakthrough moment has been awaited in NLP. Events in the last eight months strongly indicate that NLP is finally at its ImageNet moment. There are two significant advances pointing to this
Personal Digital Assistants
Imagine a world where you have a personal business assistant sitting inside your smartphone. With the intelligence to understand all your interactions and exchanges with a specific client. This would be based on not just email exchanges, calendar blocks and meeting updates, but also on external information available on the internet. Now, imagine a smart data assistant engaging with you in natural conversational English, proactively prompting you with the right insights for a client meeting, and anticipating and addressing your sales and business queries, and capturing your feedback. Rene’ ( Resonance Data ) – a smart data assistant, eliminates the need for pre-built reports.
Rene’ also gets rid of many frustrating hours spent by sales teams every week updating the CRM systems. It improves organizational productivity by freeing up valuable time of business users. Using Rene’ is just like speaking to another human being, who understands your day-to-day business conversation.