Is natural language interface to database a dead end [closed]

Question

Just want to know whether the field called the natural language interface to database is a dead end Nothing more has been developed on it since past 20 years and no latest papers are also being published.

Asking any expert in the field to shed the light

Answer 1

No, quite the contrary: today this type of technology is topical like never before. Natural Language Processing technologies are developing rapidly, and, with them, the Natural Language Interfaces. The solutions that allow to communicate with computers in natural language are crucial, because they enable easy access to data and help bridge the communication gap between humans and computers.

For example, today’s databases of corporations are so gigantic, that they can only be approached by experienced programmers. Retrieving information from a database usually requires special skills such as knowledge of SQL, however, the majority of us who interact with databases on a daily basis don’t have that background. That’s why there is a growing need for non-techie people to be able to retrieve information from databases without having to write SQL queries. And Natural Language Interfaces solve this problem.

Today, there are several NLP to database solutions that have definitely found a ready market, because they are dramatically simplifying the process of dealing with data and making data accessible for everyone! One of the popular examples of the NLP-to-database solutions is the FriendlyData API - a pioneering cloud-based NLP solution for databases, which translate natural language into query to the database. LINK

And this is going to be the new era of human-computer relations! According to Gartner, by 2020, half of analytical queries will be generated using natural language or voice-based search. And Natural Language Processing is in Gartner's TOP 10 Strategic Technology Trends in 2017.

Here is a good article about the role of NLP and the overview of the best NLP APIs & NLP-to-database APIs, where you can learn more.

Answer 2

A bit late to the party, but this is still a valid question. I'm a co-founder at DataLingvo (enterprise API to convert natural language into actions) - full disclosure...

The primary reason that you don't see much of NLP-for-SQL is that most of the modern data is NOT stored in only SQL databases anymore... The data is in myriad SaaS services with REST endpoints, NoSQL databases, Hadoop, etc. Some data is stored in SQL RDBMS still for sure but nowhere near 100% as it was 20-30 years ago.

Converting natural language interface/undersdanding (NLI/NLU) is a burgeoning area: Amazon Alexa, Siri, Cortana, Google Duplex, MS Luis, IBM Watson and DataLingvo just to name a few latest entrants.

Answer 3

No. I'll cite a few natural language query products (add http to URLs):

• DataRPM, datarpm.com/product
• Quepy (Python framework), quepy.machinalis.com/
• Oracle ATG (2010 commerce acquisition), docs.oracle.com/cd/E23507_01/Search.20073/ATGSearchQueryRef/html/s0202naturallanguagequeries01.html
• Microsoft PowerBI, support.office.com/en-ca/article/Power-BI-Q-A-in-Office-365-Searching-and-Querying-with-natural-language-709ef848-660b-4610-9b40-9395392c38af
• Wolfram natural language understanding, www.wolfram.com/natural-language-understanding/