Local language, NLP, bots and more: Emerging trends in search

hiddenNov 28, 2016 09:44:24 IST

By Radhakrishnan Srikanth

Web search is omnipresent in our lives today, whether it’s to keep track of cricket scores, check out restaurant reviews or compare car specs, web search is the first thing we turn to. With results delivered in seconds, chances are, we don’t spend any time thinking about the technological plumbing behind this instant gratification. That is what makes search so exciting! It has the simplest user experience and yet behind the scenes has the most complex algorithms processing it. Did you know that 3.9 Terra Bytes (39457 GB) of data is downloaded every second on the Internet?

Search has evolved from the time when PCs of the bygone era literally groaned while churning out the results, to data centers of machines with multi core processors, GPUs and FPGAs running sophisticated Artificial Intelligence algorithms like Deep Neural Networks. Over the last couple of years search has undergone a transformation primarily driven by users increasingly adopting mobile devices, resulting in the advent of Bots and Digital Assistants.

These new technologies and consequent changing trends are ushering new user interface patterns including natural language text and voice inputs. While users are aware that Artificial Intelligence (AI) powers these bots and digital assistants, they still tend to develop human like conversational interactions with these entities. These new user behaviors is bringing in more natural language both typed and voice based conversational queries to search endpoints i.e. users are now using sentences instead of keywords to perform their search.

As a leader in cloud and AI solutions, Microsoft is constantly innovating to keep up with the changing needs of the users. We are working towards future-proofing solutions by adapting frameworks that are robust, agile and engaging.

Natural Language Processing (NLP)

NLP is used to analyse, understand and generate structures from languages that humans naturally use while conversing. It intends to make your device more assistive, and allows for a more conversational style of user interaction, as if you were addressing another person. While the state of the art of NLP has progressed significantly over the past few years, natural conversations still pose significant challenges. Some of these challenges are overcome by using a mix of Machine Learning techniques and Knowledge Engineering. These new interfaces allow search services to become more versatile and begin to handle these conversational queries.

Local Language Search

India is a land of many languages, while the current internet user base is English educated and internet capable, with the advent of cheap data plans from telecom companies a large number of people who are only conversant in local language are poised to enter the internet age. Also more increasingly the government as well is pushing a digital agenda and wants to make services to the general public available over the internet, in their local languages.

This user trend is already visible as the amount of local language based web surfing is increasing at a steady pace. In addition to surfing, there is a lot of English language content which the local language users want to avail of as well. We here at Microsoft are working on making language translation and captioning between languages available, thereby helping play The Article needs to be attributed to Radhakrishnan Srikanth- Group Program Manager - AI and Research Division, Microsoft India Pvt Ltd a key role in enabling non-local language content being available for local language consumption.

Opening up Search APIs

It takes great hardware and smart software to provide a great user experience. With developers building new applications, search is starting to become an integral part of their offerings within their applications. Independent Bing search and Cognitive APIs help apps embed search technologies quickly and easily into apps, allowing developers provide a seamless and an enhanced experience to their users.

Bot Framework

Bots are now taking center stage. As these technologies evolve, the popular prediction is that conversations will be prevalent between people, digital assistants and bots. Given this, Microsoft is making available the Microsoft Bot Framework, a comprehensive offering that a developer can use to build and deploy high quality bots for their users, enabling users expedite tasks like booking an airplane ticket or reserving a table via their apps. These bots can connect on multiple channels (e.g. Text/SMS, Skype, Slack, Facebook Messenger, and other popular services) and can interact with users wherever they are. This framework along with associated API offerings from Microsoft will enable developers to take advantage of technologies like Machine Learning, Deep Neural Networks, Natural Language Processing, Adult and Offensive filtering, Image Processing etc., to enable new and custom bots which are not only conversational and intuitive, but also possess deep capabilities.

So if you have a plan to buy a birthday gift for a friend, a bot or digital assistant will likely help you with end to end assistance which includes help with the right gift suggestion, payment options, set delivery preferences and the ability to track your purchase till it reaches you. With just a touch or a shout, upcoming bots and digital assistants will take care of your tasks more efficiently and seamlessly.

The future of search is going to be even more exciting. It will take new forms across new endpoints and continue the traditional focus of simplifying our lives. We at Microsoft are doing just that with Bing Technologies.

The author is Group Program Manager - AI and Research Division, Microsoft India Pvt Ltd.