Cloudian Collaborates With Skymind on HyperScale Data Management Solutions

Cloudian, Inc., in enterprise object storage systems, has teamed with Skymind,Inc., the creator of the Deeplearning4j open-source deep-learning library, to create data management solutions for the hyper-scalable data sets necessary for artificial intelligence (AI) and machine learning (ML) use cases.

Under the partnership, the two companies will collaborate on integrated, enterprise-ready AI/ML solutions. Cloudian will provide the scalable on-premises storage systems that bring hyperscale technology to the data centre, while Skymind’s Deeplearning4j recognises patterns in massive amounts of data to provide actionable intelligence for companies.

Industries benefitting from the partnership will include healthcare, fraud detection for financial services, security, robotics.

Data management is an integral part of the AI/ML landscape. According to  IDC, spending on cognitive-related servers and storage will total $1.9 billion in 2017. Global spending on cognitive and AI solutions will continue to see significant corporate investment over the next several years, achieving a CAGR of 54.4%, IDC projects.¹

“Cloudian and Skymind help organisations extract actionable information from vast pools of unstructured data,” said Michael Tso, CEO and co-founder, Cloudian. “Training data is the foundation of AI and ML. More training data and better metadata tagging result in quicker, more effective learning, making data the dynamic asset that is the key to unlocking AI and ML’s business impact.“

The first solution resulting from the partnership is a security monitoring tool that identifies likely network attacks. The solution uses pattern recognition to identify suspicious network traffic. Users can see both the source and the destination of the attack, helping to simplify countermeasures. Cloudian is supplying the data management that supports the training environment, the algorithm development and its ongoing refinement.

“Deep learning can produce state-of-the-art accuracy on many problems, and it requires large datasets to do that,” said Chris Nicholson, CEO, Skymind. “That’s why it makes sense for a deep-learning company like Skymind to partner with a hyperscale storage provider like Cloudian. These are two crucial parts of a larger solution.” 

Organisations are poised for rapid adoption of AI and ML as Skymind and other suppliers develop the toolkits and support expertise that make technology accessible. With AI and ML growth, however, storage challenges follow, as organisations manage massive and rapidly growing training data.

The utility of Cloudian’s hyperscale storage for AI was previously demonstrated in 2016, when the company partnered with Dentsu, Inc. and Intel Corp. to launch a digital signage system with advertisements targeting specific demographics via machine learning. As cars approached the roadside LED billboards, Cloudian’s HyperStore technology allowed Dentsu to change advertisements in real time based on the year, make and model of passing vehicles.

Skymind’s technology already has demonstrated its power. The municipal government of Fuzhou, China, is working with Skymind to help predict heart disease among its citizens using ultrasound videos with 80EB of data. The project included the construction of a new data centre as well as a large GPU cluster.

Skymind also has helped several major banks and telecoms build fast fraud detection solutions. Its neural net model server returns results in nanoseconds, and, with the proper dataset to train on, its algorithms can attain human-level accuracy in their predictions.

¹ IDC Press Release, April 3, 2017