NLP - the speaking Avaatar of Artificial Intelligence

Date:   Tuesday , January 31, 2017

SAP (NYSE: SAP) is a German multi-national software development company developing enterprise software solutions that enable businesses to manage their operations and customer relations in an efficient and hassle-free way. It has a current market capital of $108.06 billion.

Last recorded, India stood at a population of 1.252 billion. 22 official languages, 1652 mother tongues, and a gazillion dialects make India perhaps linguistically the most diverse country in the world. And while we talk of jumping sharks over land and flying cows through blue skies with artificial intelligence, here\'s an avaatar that can actually make people talk to each other. Not like a Facebook talk to each other. But like a Punjabi, Hebrew and Bengali chat with each other despite each talking in their mother tongue! Sounds like gibberish? But that\'s exactly the kind of impact that NLP can affect. And it can do much more than just get people to chat.

But First, What Is NLP & Why Are People So Excited About It?

NLP exploration on our current digital planet has impacted in various forms in the voice to speech mode, in the form text to speech. The resultant output has been the voice searches on automobiles, and then of course the dictation mechanics of the software world. It became a popular framework when Apple introduced SIRI and Google introduced the YouTube voice search. What does it do? It took over the popular demographics amidst the technology enthusiasts and the kids community using SIRI.

Does NLP Only Help Mankind Cosmetically?

SIRI has cracked the most popular use case for language processing via NLP. Apple\'s NLP on voice to text works on 16 global languages and they are meant for languages from English to French to Chinese. These have had an impact on the iPhone community when it comes to uses of searches, or looking for your content in your phone, and has made life easier, lazier and efficient. Multi-lingual internationalization exists for YouTube as well. However, NLP applications can extend into fields with deep reaching impact.

Use cases of NLP vary from SIRI kind of applications to more commercially viable ones. Realty will most likely be the highest and quickest adopter of NLP, supporting the homes with on-voice command executions. Yet again, there are several unexplored aspects of this division of NLP that seek attention -

  • Medical Aid: NLP can actually make gadgets and homes for physically challenged a cake walk. It could make the challenges and costs of affording an expensive on-demand nurse less necessary

  • Education: By enabling speech to text conversion, NLP has the opportunity of reducing language barriers. This in itself will overcome a significant challenge considering that 90 percent of human interactions are driven by language and communications

  • Business Decision Making: NLP and NLG together can actually analyze and present high volumes of data enabling simpler and more efficient decision making

  • The Macro Perspective

    Let\'s take a step back and look at things from a different perspective. Let\'s explore the example of the now popular Terminator Age. Prima Facie, the conversation revolves around the potential impact of a machine learning process driven by software and neural modeling handling gazillion data packets. Now what use are data packets? Where do they take us? Why do this entire thing at all?

    A simple use case demonstrates the potential - over 800 million Indians do not even have access to hospitals and clinics with sufficient hardware. The impact is not just in their healthcare opportunities but also the loss of extraordinary amounts of data & knowledge that in times of epidemics could prove to be game changers for the country. Upon applying NLG and NLP, the use case becomes even more robust. In a Punjabi village, a Punjabi speaking doctor can now converse with a Hebrew speaking specialist residing in Israel to resolve a serious medical issue in the back and the beyond. None of this would even be possible without machine learning and application of NLP.

    Critical data on symptoms, diseases and chronic conditions accumulate continuously and are neither being collected nor used constructively to lead to any potential short term benefit - leave alone the long term benefits. Consider this use case - A simple machine learning algorithm can analyze, predict and recommend the MRI scanners while the scan of a patient is going on! This will land up fast tracking the process of remedies the doctors or the clinics can recommend. In times of crisis, this simple measure can be the difference between life and death of a patient!

    India & NLP

    I am often asked this question when talking about NLP - Why do I feel India can actually crack the NLP model? Of all the divisions of artificial intelligence, I personally believe that the next generation evolution in NLP will come from India. It brings me to what I fundamentally believe; if a country with 1652 mother tongues and multiple other dialects cannot crack the NLP model, then perhaps few others can. And work is already underway with more and more talented Indians setting their eyes on it - like MIHUP. India has always been a research hub of NLP, and organizations such as Owler are great, but the mantle to crack NLP for 22 languages by voice to text, and augment voice commands to machines via key lexicon datasets on universal remotes and many more use cases is taken by an organization called MIHUP. Funded by Accel and backed by a diverse team with a deep understanding of language and linguistics, MIHUP is able to crack 22 languages and ensures that machines and device commands are getting augmented through their algorithms, then India leap frogs hugely as a subcontinent on NLP innovation.

    What Can We Expect This To Do For Society?

    For instance, uneducated farmers talking on platforms on crop remedy, tackling farmer suicides, medical aid or support, and many more never ending opportunities exist. The idea is to identify them, identify barriers, and then work on creating a robust model; one that bridges back to society and social benefit - an objective that all players in the NLP field should aim for with good reason. After all, a technology that focuses on a 90 percent consumption pattern should be the first one to actually carve that path and ensure that humankind keeps talking to each other!