Natural Language Processing is mainly about bridging the gap between how humans communicate and what computers understand. Can machines really understand what humans talk? Yes! By combining the power of artificial intelligence, computational linguistics, and computer science, NLP allows a machine to understand natural language. A task that, up until now, only humans could perform.
When you are engaged in a conversation with a person, each of you understand the meaning of what the other person is saying. On the other hand, computers understand only machine language, and are programmed to understand very specific instructions on top of this language and hence cannot comprehend the ambiguity of the vast human language.
Human knowledge is limited, while today, in the big data era, the computer has access to almost unlimited knowledge.
So what if we taught computers to understand us?
Why?
Today’s machines can analyze more language-based data than humans, without fatigue and in a consistent, unbiased way. Considering the staggering amount of unstructured data that’s generated every day, from medical records to social media, automation will be critical to fully analyze text and speech data efficiently. Natural language processing helps computers communicate with humans in their own language and scales other language-related tasks.
For example, NLP makes it possible for computers to read text, hear speech, interpret it, measure sentiment and determine which parts are important.
Natural language processing helps computers communicate with humans in their own language and scales other language-related tasks. For example, NLP makes it possible for computers to read text, hear speech, interpret it, measure sentiment and determine which parts are important.
Use Cases
NLP is everywhere, even if we don’t know it. Although the term is not as popular as Big Data or Machine learning, we use NLP every day. Maybe you have already used machine translation and it seems a natural feature to you by now. The globe icon on Twitter or the translate links in Facebook posts, in Google and Bing search results, in some forums or user review systems are classic examples.
“Hey, Siri! Call Mom”
“Alexa! Reduce the volume”
Did you know that voice assistants are the classic examples of NLP?
Amazon’s Alexa and Apple’s Siri are examples of intelligent voice driven interfaces that use NLP to respond to vocal prompts and do everything like find a particular shop, tell us the weather forecast, suggest the best route to the office or turn on the lights at home.
Did you know that NLP can be used in predicting diseases?
Amazon Comprehend Medical is a service that uses NLP to extract disease conditions, medications and treatment outcomes from patient notes, clinical trial reports and other electronic health records. Stanford University developed “Woebot”, a chatbot therapist with the aim of helping people with anxiety and other disorders.
Have you ever looked at the emails in your spam folder and noticed similarities in the subject lines?
You’re seeing Bayesian spam filtering, a statistical NLP technique that compares the words in spam to valid emails to identify junk mail.
Have you ever missed a phone call and read the automatic transcript of the voicemail in your email inbox or smartphone app?
That’s speech-to-text conversion, an NLP capability.
Ever navigated through a website by using its built-in search bar, or by selecting suggested topic, entity or category tags?
Then you’ve used NLP methods for search, topic modelling, entity extraction and content categorization.
NLP is an applied artificial intelligence (AI) program that helps your chatbot analyze and understand the natural human language communicated with your customers. Chatbots are able to understand the intent of the conversation rather than just use the information to communicate and respond to queries.
Text Analysis
Most of us struggle to understand the tone of someone’s speech or the sentiment behind it.
The biggest irony is that we can even describe ourselves as a more evolved version of a computer!
While we struggle with day to day conversations, a machine which is based on 1’s and 0’s will not be able to differentiate cat and dog.
We need to teach it a method of thinking. Here are some of the techniques it can be taught:
Syntactic Analysis
Analyses of text using basic grammar rules to identify structure, how words are organized, and how the words relate to each other.
1.Tokenization :-
The process of breaking the raw text into chunks of words that will be processed later. These chunks are called as tokens. It helps in finding out the meaning of the phrase by looking at the sequence.
E.g. “I love potatoes” will be broken as ‘I’, ‘love’, ‘potatoes’.
2. Part of speech tagging (POS tagging) :-
The process of labelling tokens as verb, adverb, adjective, noun, etc. This helps in finding out the meaning of a word and will give insight into the different contexts with which the word can be used.
3. Lemmatization & stemming :-
The process of reducing the selected words to their base form to make them easier to analyze. When we get to the root of a particular word, we can easily find out what it means. Stemming chops off the unwanted words at the end of the word for it to make sense and lemmatization retrieves finds the base (or root) word.
E.g. Stemming will reduce the word ‘biscuits’ to ‘biscuit’, by chopping off ‘s’ at the end.
Lemmatization can match the word biscuits to ‘biscuit’ and ‘food’.
4. Removing Stop-words :-
The process of removing frequently occurring words that don’t add any semantic value. Certain words like I, they, have, like, the, yours, etc., don’t add much meaning to the sentence except connect it to each other.
They do add meaning to the sentence, but while analyzing the semantics of the entire context, their worth falls short.
Semantic analysis
Analyses the meaning of text. First, it studies the meaning of each individual word, then it looks at the combination of words and what they mean together.
1.Word sense disambiguation :-
The process of determining which meaning of the word is activated while using that word.
E.g. “Let’s go play in the park.” and “The play we saw today was absolutely marvelous!” are two sentences with the same word but different meanings.
2. Relationship extraction :-
The process of understanding how words (places, persons, organizations, etc) relate to each other in a text.
E.g. “Tamil Nadu is in India” , in the given sentence ‘Tamil Nadu’ will be categorized as city and ‘India’ will be categorized as a country.
Sentiment Analysis
Sentiment Analysis is the process of extracting opinions from a specific piece of text written in natural language (the language with which we talk and write) using Natural Language Processing (NLP), text analysis and statistics to analyze the individual’s sentiment.
In other words, it finds out whether the particular piece of text is positive, negative, or neutral, thus revealing the sentiment (the subjective meaning) of the text conveyed.
By analyzing social media posts, product reviews, or online surveys, companies can gain insight into how customers/ consumers feel about their brands or products.
A Sentiment Analysis model is taught Natural Language Processing techniques to spot positive and negative keywords.
Where is it used?
Politics:
Political parties and campaign managers use sentiment analysis to capture the opinion of people on specific topics.
Twitter sentiment analysis has helped political parties plan out their campaigns and strategies. It helps them understand the opinion of the general public efficiently. We can take the data from the public’s tweets on a specific topic and analyze it to see whether the response was great or not.
Customer Experience:
Companies can check the reviews of a particular product or service as well as the opinion of their customers to see if they like it or not. Sentiment analysis plays a powerful role here.
Voice of customers is vital for any company. After finding out the opinion of the customers, the organization can come to a conclusion if it needs to improve its product or not.
If a product isn’t getting a positive response, the company would stop selling it or improve it. All of this helps enhance the customer’s experience.
Competitive Analysis:
Companies can also analyze it’s competitor’s content to find out what works with the public that you may not have considered. It will help the company with benchmarking.
The company can understand it’s strengths and weaknesses and how it matches up to its competitors. Sentiment Analysis is a powerful tool that is used here.
Conclusion
This is just the beginning of what humanity can do with its growing need for change and luxury. But no matter what improvement we have, emotions and our sentiment will not change according to the times. Natural Language processing is a powerful subfield of AI that can help us achieve the epitome of our very existence — machine with a mind.
This was just a quick glance of what we can do by marrying the nature of man to the chips of a computer.
Take a peek before you start your journey into the ocean!
By:- Ramya Kalyanasundaram and Swetha Ramachandran
