Home General IBM Unveils Deep Learning as a Service (DLaaS) for AI Developers

IBM Unveils Deep Learning as a Service (DLaaS) for AI Developers

During its Think 2018 conference recently held in Las Vegas, IBM unveiled its Deep Learning as a Service program meant for artificial intelligence developers. The service can be accessed through Watson Studio.

The aim of creating the Deep Learning as a Service (DLaaS) program is to allow AI developers to run hundreds of deep learning models simultaneously while still building their neural networks.

Thanks to this program, developers are now in a position to use popular frameworks such as PyTorch, Caffe2, and TensorFlow among others to train their deep neural networks.

The advantage of IBM’s DLaaS program is that users only pay for the GPU time while using only the resources that they require for training their models.

The service saves a significant amount of time for developers. In fact, they only need to clean their data, upload it, commence training before they can download and observe the training outcomes.

According to IBM, the service can take days or weeks to process iterative training models.

According to a White paper written by IBM researchers, IBM has been developing the service since mid-last year.

By using the power of the cloud to provide AI capabilities such as deep learning, the company and other vendors with similar services are democratizing access to their tools.

Thanks to such services, companies do not have to create and maintain expensive hardware in an attempt to explore deep learning. As such, additional companies could utilize the power of artificial intelligence in developing their products and services.

The white paper also revealed that the Deep Learning as a Service (DLaaS) product comes with a resources provisioning layer. It enables the managing of jobs across heterogeneous resources, for instance, the use of both central processing units and graphics processing units in running workloads.

Upon collecting, organizing and uploading the data to the service, IBM automatically begins the training and stops once it is done. This process helps in provisioning resources to protect the users from overpaying.

Aside from the DLaaS program, IBM also revealed details on the new IBM Power Servers, which are designed for AI.

It said that the new servers would be moved to various cloud data centers. Furthermore, the release indicated that the IBM Cloud Object Storage would now also include a high-data transfer option dubbed IBM Aspera for transferring data to the cloud.

During the Think conference, IBM also revealed great news for the company citing that its Hyper Protect Family of services had attained FIPS 140-2 Level 4 certification.

The release also highlighted new features for the IBM Cloud Private platform, which allows businesses to develop their clouds in either external or internal data centers.

The launch of DLaaS has made IBM join the ranks of other companies that are trying to simplify the training of deep neural network models. Some of these entities include Amazon, Google, and Facebook.

Identifying experienced machine learning practitioners and data scientists is both expensive and challenging for organizations.

As such, programs or services like Google’s AutoML and IBM’s DLaaS are making machine learning as independent as possible.

Source Analytics

Subscribe to our newsletter

Signup today for free and be the first to get notified on the latest news and insights on artificial intelligence

KC Cheung
KC Cheung
KC Cheung has over 18 years experience in the technology industry including media, payments, and software and has a keen interest in artificial intelligence, machine learning, deep learning, neural networks and its applications in business. Over the years he has worked with some of the leading technology companies, building and growing dynamic teams in a fast moving international environment.
- Advertisment -


How can AI be leveraged for Clinical Trial Prediction?

Successful clinical trials occur at a ratio of 1 in 10, costing around $2-$3 billion, with drugs taking 10–12 years to be approved. High costs,...