IJCAI 2019: Hike Research Paper On AI And ML Gets Published

Hike, a homegrown internet startup has unveiled its research paper at IJCAI (International Joint Conferences on Artificial Intelligence) 2019. Its a non-profit organization which allows companies and researchers to showcase their researches. Hike became one of the few Indian companies to get shortlisted and present its research paper that focuses on the role of AI integration in Hike Sticker Chat for sticker recommendation.

We at Gizbot got the opportunity to interact with Ankur Narang, VP of AI and Data Technologies at Hike and Mohamed Hanoosh, one of the authors of the paper, to know more about it. Here's what they have to about this new feat:

What makes the paper submission unique with regards to sticker recommendations?

According to the spokespersons, the paper submitted at IJCAI is unique because it addresses the challenges which developers face with the sticker recommendation. The paper explains the algorithmic challenges occurs in understanding the chat and in giving sticker recommendations. It also explains system challenges to enable low-end devices with good quality recommendations

"The paper gives us a deep insight into the phenomenon of how sticker recommendations work in the application and how that AI and ML can be used to find effective solutions. It further elaborates how Hike is working on NLP in local languages that give us a unique advantage," Narang commented.

How will this impact the end-user?

Hike Sticker Chat has the advantage of using conversational stickers. Since it was launched back in April 2019, the online messaging app has more than 1 million weekly active users using more than 60,000 stickers. Hike stickers allow users to chat with their friends and family without typing much in a fun and exciting way.

Through the findings of this paper, "we aim to make sticker led conversations between users even more seamless. Our algorithm will play a key role in helping find the right stickers at the right time with minimum effort making Hike Sticker Chat a great place for young Indians to express themselves with their close friends," Hanoosh added.

Can you explain how AI and ML work in suggesting stickers?

According to the VP and author of the paper, the role of AI and ML becomes extremely critical in suggesting the right stickers at the right time. In Hike Sticker Chat, developers are bringing Machine Learning, infinite content creation, and voice input all together.

"Our ultimate mission is to reduce people's dependency on the keyboard- type less and express more; this can be accomplished by using the right mix of AI and ML in sticker recommendations," they added.

What's The Company's Focus on Research and Development?

According to the spokesperson, the company is working closely with the ecosystem to enable the field of AI and ML. Hike also believes in multiple collaborations across academic disciplines that can leverage the bright minds of the country. Recently, the company has also published a paper on Friend "Recommendation" at ECIR 2019. Out of 151 submission Hike's paper got selected.

"Today, we are the only player in the market to be solving problems at such massive regional volume with a focused inclusive and hyper-local approach catering to low-end handset & low bandwidth market areas," Hanoosh added

How does this make Hike Sticker Chat different from other social media platforms?

India has more than 22+ languages and 700 + dialects, and Hike Sticker Chat is the platform that has integrated more than 60,000 stickers in over 40 languages and dialects.

"We understand Indians and communication like no one else. Built with AI and ML at its core, we aim to make conversations simple and expressive by suggesting the right sticker at the right time," Narang added in the interaction.

What are your plans in the future?

The company will work constantly towards scaling the app by reducing the dependency on the keyboard. The company will be working more on AI and ML team to make it more accurate and innovative for users.