H2O.ai and Nvidia today announced that they have partnered to take machine learning and deep learning algorithms to the enterprise through deals with Nvidia's graphics processing units (GPUs). Mountain View, Calif.-based H20.ai has created AI software that enables customers to train machine learning and deep learning models up to 75 times faster than conventional central processing unit (CPU) solutions. H2O.ai is also a founding member of the GPU Open Analytics initiative that aims to create an open framework for data science on GPUs. As part of the initiative, H2O.ai's GPU edition machine learning algorithms are compatible with the GPU Data Frame, the open in-GPU-memory data frame.

IBM on Wednesday is announcing significant updates to PowerAI, its deep learning software distribution package, making it faster for data scientists to deploy deep learning models and easier for developers to integrate computer vision into their applications. Analysts say that artificial intelligence has reached a tipping point where it's being integrated into just about every service, product, or integration, but there are still major challenges for the data scientists and developers interested in exploiting AI. Some sectors like financial services have had data scientists on staff for at least five or 10 years, but they've only recently started deploying deep learning methods. PowerAI "gives them these higher level tools that much it make easier and automated," IBM VP Sumit Gupta told ZDNet. "You still have data scientists guiding the whole process, but we're removing some of the steps."

HPC blog Just as Intel, the king of CPUs and the very bloodstream of computing announced that it is ending its Intel Development Forum (IDF) annual event, this week in San Jose, NVIDIA, the king of GPUs and the fuel of Artificial Intelligence is holding its biggest GPU Technology Conference (GTC) annual event yet. With something north of 95 per cent market share in laptops, desktops, and servers, Intel-the-company is far from even looking weak. Indeed, it is systematically adding to its strengths with strong indigenous high-density processing CPUs of its own, acquisition of budding AI chip vendors, pushing on storage-class memory, and advanced interconnects. But a revolution is nevertheless afoot. The end of CPU-dominated computing is upon us.

You hear a lot of different terms bandied about these days when it comes to new data processing techniques. One person says they're using machine learning, while another calls it artificial intelligence. Still others may claim to be doing deep learning, while "cognitive" is the favored phrase for some. What does it all mean? While many of these terms are related and can overlap in some ways, there are key differences that can be important, and that could be a barrier to fully understanding what people mean when they use these words (assuming they're using them correctly).

The role of social media, in particular microblogging platforms such as Twitter, as a conduit for actionable and tactical information during disasters is increasingly acknowledged. However, time-critical analysis of big crisis data on social media streams brings challenges to machine learning techniques, especially the ones that use supervised learning. The scarcity of labeled data, particularly in the early hours of a crisis, delays the learning process. Existing classification methods require a significant amount of labeled data specific to a particular event for training plus a lot of feature engineering to achieve best results. In this work, we introduce neural network based classification methods for identifying useful tweets during a crisis situation. At the onset of a disaster when no labeled data is available, our proposed method makes the best use of the out-of-event data and achieves good results.

Neural networks require a careful design in order to perform properly on a given task. In particular, selecting a good activation function (possibly in a data-dependent fashion) is a crucial step, which remains an open problem in the research community. Despite a large amount of investigations, most current implementations simply select one fixed function from a small set of candidates, which is not adapted during training, and is shared among all neurons throughout the different layers. However, neither two of these assumptions can be supposed optimal in practice. In this paper, we present a principled way to have data-dependent adaptation of the activation functions, which is performed independently for each neuron. This is achieved by leveraging over past and present advances on cubic spline interpolation, allowing for local adaptation of the functions around their regions of use. The resulting algorithm is relatively cheap to implement, and overfitting is counterbalanced by the inclusion of a novel damping criterion, which penalizes unwanted oscillations from a predefined shape.

Research into artificial intelligence is going gangbusters, and the frenetic pace won't let up for about five years -- after which the industry will concentrate around a handful of core technologies and leaders, the head of Intel's new AI division predicts. Intel is keen to be among them. In March, it formed an Artificial Intelligence Products Group headed by Naveen Rao. He previously was CEO of Nervana Systems, a deep-learning startup Intel acquired in 2016. Rao sees the industry moving at breakneck speed.

Jensen Huang beamed out over a packed conference hall in San Jose, California, on Wednesday as he announced his company's new chip aimed at accelerating artificial intelligence algorithms. But metaphorically speaking, the CEO of chip maker Nvidia was looking over his shoulder. Nvidia's profits and stock have surged over the past few years because the graphics processors it invented to power gaming and graphics production have enabled many recent breakthroughs in machine learning (see "10 Breakthrough Technologies 2013: Deep Learning"). But as investment in artificial intelligence soars, Huang's company now faces competition from Intel, Google, and others working on their own AI chips. At Nvidia's annual developer conference on Wednesday, Huang carefully avoided mentioning any competitors by name as he introduced Nvidia's latest chip, named the Tesla V100.

So, you want to create a hugely successful machine-learning startup? Or you've been asked to start investigating ML for your firm? Well, you'd better get programming – but what language should you use? No languages have been designed specifically with ML in mind, but some do lend themselves to the task. Developers experimenting with machine learning will spend most of their time processing data sets, running them against a machine-learning algorithm, and then classifying them again until the results seem right.

If you've been waiting for NVIDIA to finally take the lid off of Volta, the next generation of its GPU technology, your day has finally come. Today at its GPU Technology Conference, the company announced the NVIDIA Tesla V100 data center GPU, the first processor to use its seventh-generation architecture. Like the Tesla P100 the processor is replacing, the Volta-powered GPU is designed specifically to power artificial intelligence and deep learning so, naturally, it's flush with power. Built on a 12nm process, the V100 boasts 5,120 CUDA Cores, 16GB of HBM2 memory, an updated NVLink 2.0 interface and is capable of a staggering 15 teraflops of computational power. Naturally, it's also the GPU that drives the company's updated DGX-1 supercomputer, too.