FOLLOW

FOLLOW

SHARE

Tech Giants Rush To Capture AI Startups

How the tables have turned, with large companies chasing startup talents, instead of being chased by them

5Jul

AI is no longer a subject for discussion amongst futurists, it's a reality that creates many opportunities for investors. According to Internet Live Stats, around 40% of the world's population is online which means that the world operates with enormous amounts of data, and AI technology sees it as a raw material to work with. Developments in deep learning have helped to find out more about our behaviors, interests, knowledge, and connections. Tech giants are not missing out and want to use this knowledge to improve and be ahead with the latest innovation. Whilst large companies have labs and incubators to test the latest technologies, they also keep an eye on the startup community to foster innovation.

One of Google’s first investments in AI was a Canadian startup, DNNresearch, which they acquired in 2013. The startup's name stands for 'deep neural networks', and it aims to ease 'training' processes for artificially intelligent systems by implementing an object recognition technology based on neural networks. DNNresearch is led by Geoffrey Hinton, a professor at the University of Toronto, and two of his graduates who developed an AI system based on deep learning and convolutional neural networks. The system helps to make photo search more precise, so previously, if a person wanted to search, let’s say, an image of a specific flower by its name, the search would show images of that flower, but would also include similar flowers or irrelevant images that are associated with the word in the search box. By using neural networks, Google has managed to solve the problem, plus the company got Geoffrey Hinton on board as one of its AI specialists.

Another step that Google made towards improvement of its AI strategy was the acquisition of DeepMind. The deal was believed to capture new scientific thinkers, rather than specific products or services. Yoshua Bengio, an AI researcher at the University of Montreal told MIT technology review that the main reason Google has acquired DeepMind is because it has the largest concentration of deep learning experts. According to Forbes, Facebook had also tried to get to recruit DeepMind and its bright founder, neuroscientist Demis Hassabis alongside his team, but failed. For Google, the deal helped to achieve a major breakthrough after the DeepMind team taught a computer the ancient game of Go, considered a big challenge for AI to learn. Facebook, didn't accept the AI victory, with Yann LeCunn, Facebook's Director of AI research, commenting: 'It wasn't true artificial intelligence.'

Another startup gem that is set to foster AI innovation, but this time for Twitter, is Magic Pony, acquired in June 2016. It's the third machine learning startup purchase for the company, alongside Whetlab, bought in 2015, and Madbits, in 2014. The deal was worth $150 million, indicating that Twitter has serious intentions on succeeding in the AI race. Jack Dorsey, Twitter CEO and co-founder, once said, 'Machine learning is increasingly at the core of everything we build at Twitter.' The startup was founded by Imperial College London graduates, Rob Bishop, and Zehan Wang. Magic Pony uses systems that essentially work like human brains (known as neural networks) and machine learning to improve the features of imagery, by enhancing quality, particularly in videos captured on mobile phones, that may lack quality because of the poor filming conditions. At the moment, social media platforms compress large size videos in order to stream them, which results in loss of quality. The tool is of paramount importance for social media platforms like Snapchat, Facebook, and Twitter.

CBInsights's data shows that 2016 is set to become a new high in the AI deals and acquisitions, with 24 deals made in the Q4'15, and Q1'16 has already reached a 5-year quarterly high, passing the 25 deal threshold. 

Comments

comments powered byDisqus
Channelsspare5

Read next:

2016 State of Crowdsourced Innovation

i