Qihoo 360, a Chinese tech company best known for its anti-malware software, is turning over a new leaf in product development.

Following the lead of other Chinese tech giants, Qihoo 360 launched its own AI research institute and is looking at developing smart driving applications.

“We have a very clear and long term target,” Shuicheng Yan, the Chief Scientist at Qihoo 360’s AI research institute, told TechNode. “Definitely it’s for smart driving. […] From product’s perspective, I consider smart driving as a major focus on the whole research institute.”

Leveraging Dr. Yan’s academic background in computer vision and deep learning, Qihoo 360’s AI research institute will primarily focus on image and facial recognition. While strengthening Qihoo 360’s existing IoT portfolio is the institute’s priority, Dr. Yan’s team is also looking at using AI to improve driver safety.

Qihoo 360 will start at the “zero level” of autonomous driving, he says, with advanced driver assistance systems (ADAS), including a rear-view mirror that helps people drive more safely. The company also plans to develop products that monitor driver behavior and assess the environment around the vehicle.

The company is also taking advantage of their own strengths in security.”If you have various connections within the car [or] if you want to connect [your entertainment system] to the internet, definitely you will have the security threatened,” says Dr. Yan.

Qihoo 360 plans to develop security software reminiscent of the company’s “Safety Bodyguard” [our translation] anti-malware mobile app. However, whether or not Qihoo 360 will go as far as to develop their own fleet of autonomous cars, similar to that of Baidu or Google, is still under discussion. The company also plans to conduct their own research on voice and speech recognition, but under a separate research organization, says Dr. Yan.

Pivoting To IoT

Qihoo 360’s smart driving plans are part of the company’s overall goal to focus on connected devices. At the Second World Internet Conference last December, Zhou Hongyi, chairman and CEO of Qihoo 360, dubbed IoT the best business opportunity in the next five years. In many ways, the company’s AI research institute will be an extension of its IoT product development unit.

“We mainly support the two major lines of products of the company,” says Dr. Yan. “One is smart devices, IoT. Another line is the livestream[ing].”

Dr. Yan’s team is improving the facial recognition features of Qihoo 360’s smart home security camera, “Small Water Droplet” (小水滴, our translation). For Qihoo 360’s livestreaming platform, Huajiao (花椒), the research institute will enhance face tracking features, such as beauty and face swapping filters. At the moment, fundamental research is not a priority, says Dr. Yan.

IoT and AI could generate new revenue streams for Qihoo 360, whose main source of revenue comes from advertising on platforms like 360 Search and 360 Mobile Assistant, Qihoo 360’s mobile app store. Last year, online advertising services accounted for 67.1% of the company’s total revenue. In contrast, revenue from smart hardware and IoT devices was about 3% of Qihoo 360’s 2015 revenue, 88% of which was cost.

Qihoo 360 will also face steep competition from more established players. Other domestic tech giants, such as Alibaba and Baidu, started investing in AI years ago, either through partnerships, such as Alibaba’s partnership with facial recognition company Face++, or their own proprietary research labs, such as Baidu’s Institute of Deep Learning.

As the new kid on the block, Qihoo 360 will not only have to boost AI capabilities of existing products to survive, but develop cutting edge applications of its own.

Disclaimer: Although Qihoo 360 provided no editorial control over this post, the company covered the travel expenses involved in interviewing Dr. Shuicheng Yan.

Eva Xiao is a tech reporter based in Shanghai. Contact her at eva.xiao@technode.com or evawxiao (wechat & twitter).

Leave a comment

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.