6 January 2014


3 mins read

NLP, Wearable Tech and Voice Search

Maratopia Digital Marketing Ltd

NLP, Wearable Tech and Voice Search

What is Natural Language Processing (NLP)?

Natural language processing (NLP) is the ability of computer programmes to understand human speech, as it is actually spoken. That means NLP has to tackle the often ambiguous and highly complex linguistic structures people use in everyday speech. As such there are many variables these computer programmes have to understand such as: slang, errors, regional dialect and social context, in order to process language correctly and indeed, naturally. Typical approaches to NLP are based on machine learning, which is a type of artificial intelligence centred on identifying the uses and patterns in data.

Most of today’s NLP research revolves around search.

Ultimately the task for NLP is to eliminate the need for computer programming languages such as Java, PHP and ColdFusion. Instead, if NLP is successful, we will simply communicate with machines in “human” languages. However, when it comes to achieving this task there are quite a few challenges that need to be overcome, challenges including:

  • Meeting the expectations of the users
  • Understanding ambiguity in natural language
  • Understanding the effect of context on meaning
  • Understanding anaphoric referencing
  • Speed and efficiency of the interface
  • Recognizing relevant data for NLP will disregarding irrelevant data in the input like age and gender.

Generally speaking NLP has been successful in handling the challenges posed by the syntax (structure) of natural language, but researchers and programmers still have a long way to go before they meet the challenges posed by the semantics (sense and meaning) of natural languages. The main issues to solve are: understanding the meaning of a single word, understanding the meaning of that word in connection with other words in the syntax and finally understanding both those meanings in the context in which they are spoken.

Some of the these contexts in which utterances are spoken are: time, place, situation…


Uses and Applications of NLP

Siri is Apple’s almost infamous personal assistant for the iOS operating system. Siri uses NLP to answer questions and make recommendations.

Google Now is essentially Google’s answer to Siri, it is another personal assistant that uses NLP to answer questions, make recommendations and perform actions. It was named the innovation of the year in 2012.

Finally Google Glass is perhaps the newest and most widely reported on piece of technology that uses NLP. It different from Siri and Google Now in that the technology is embedded in a wearable device so the majority of commands that operate glass are delivered verbally. It is arguably the most advanced application of NLP to date though for how long that remains to be seen. You can read more about what Google Glass actually does and the role NLP plays in that here.

Google’s NLP Research

Google is working on processing multiple natural languages at web scale, and they aim to do this by leveraging the large amounts of data they have access to. In true Google fashion this involves writing algorithms to predict things like: the part of speech tags for each word in a sentence and the various relationships between them. This handles the syntax of language. To handle the semantics Google is working on solving problems like identifying noun phrases in free text and what they refer to. They do this in free text, across documents and against a knowledge base.

Microsoft’s NLP Research

Microsoft is aiming to tackle NLP using a combination of knowledge engineered and statistical machine learning techniques to remove the ambiguity in natural language. The implications of this work are far reaching and could have an impact on applications for “text critiquing, information retrieval (search), question answering, summarisation, gaming and translation.” In fact Microsoft’s NLP research and progress is already in use in many of their products such as the grammar checkers in office, Encarta, Intellishrink and the Microsoft Product Report.

(N.B. we’d have liked to have told you a bit more about how Apple plans on improving their NLP techniques, but as you can imagine, that information is as rare as fairy dust).


The Future of Search and Technology

Search is most definitely going to move away from structured, keyword based search queries that search engines interpret using algorithms, towards more conversational and unstructured search queries.

The implications are that hands free technology could really being to dominate the search market by making voice search truly effective. Products like Apple’s iPhones and Google’s Glass could begin to replace other technologies that do not offer conversational, voice search. This means that the primary way we interact with technology is developing and changing and therefore so is the way we search and discover information.

NLP also has real benefit for end users as it will eliminate the need to formulate appropriate search queries in order to return the results you want, instead you will simple be in conversation with technology.

One major barrier between man and machine will be broken.

You can also find our thoughts on these subjects below, in our latest Slideshare, Natural Language Processing (NLP), Search and Wearable Technology from Cloudspotting

In need of fresh content?