In Machine Learning, Cross-validation is a resampling method used for model evaluation to avoid testing a model on the same dataset on which it was trained. This is a common mistake, especially that a separate testing dataset is not always available. However, this usually leads to inaccurate performance measures (as the model will have an almost perfect score since it is being tested on the same data it was trained on). To avoid this kind of mistakes, cross validation is usually preferred. The concept of cross-validation is actually simple: Instead of using the whole dataset to train and then test on same data, we could randomly divide our data into training and testing datasets.

You didn't have to be following the coverage from this year's CES too closely to notice one name repeatedly cropping up, across what felt like every announcement at the show: Alexa. Alexa is an artificial intelligence, a virtual personal assistant developed by Amazon. And right now, it looks likely to be remembered as AI's first major foray into the mainstream. Developed by Amazon's Lab126 R&D division, and building on the eCommerce giant's acquisition of Cambridge AI startup Evi in 2012, Alexa made its debut – there's a temptation to say'her debut', given the gendered name and voice, but let's stay neutral for the moment – back in November 2014. It was first sold as part of the Amazon Echo, a smart speaker that enabled users to ask questions, select and listen to music and, of course, order shopping using just their voice. It wasn't until last year, however, that Alexa made a real splash.

That's the year in which artificial intelligence will be able to perform any intellectual task a human can perform, according to one survey of experts at a recent AI conference. Anything and everything any person has ever done in all of history -- all of it doable, by 2050, by intelligent machines. But what can AI do today? How close are we to that all-powerful machine intelligence? I wanted to know, but couldn't find a list of AI's achievements to date.

Facial recognition is nothing new, however this is now extending beyond humans into the world of domestic cattle. Whilst'smart' cattle monitoring is more commonplace, existing systems largely require the use of physical tracking devices. Facial recognition technology will eliminate the stress of fitting these devices, allowing easy monitoring of an entire heard with minimal interaction. This is set to enable individual monitoring of group behaviour, early detection of lameness and accurate recording of feeding habits. Although hailed as the future of farming, the extent to which AI will change the daily operations of the traditional family farm is yet to be seen.

This article originally appeared on Real Money on March 6, 2017. Given how much several tech giants throw around the terms "artificial intelligence," "machine learning" and "deep learning," it's easy to assume that they're all trying to do similar things. And in some cases, the companies are indeed using AI to tackle problems that have much in common, particularly when it comes to making sense out of real-life sounds, images and dialogue. But a closer look shows that tech giants are at times not only tackling very different challenges, but employing different AI techniques to do so. IBM (IBM), which has reportedly set a goal of obtaining $10 billion in annual revenue related to its Watson AI platform by 2023, is a particular outlier -- not seen as a leader when it comes to certain large-scale, consumer-focused AI solutions, but very much one when it comes to AI offerings that require a measure of industry-specific expertise.

When you hear "artificial intelligence" or "machine learning," what comes to mind? A complicated technology that demands deep domain experience or a degree to use? This was once the way technology worked; only a select few had access. But innovation has a funny way of changing things. What might seem out of reach today can become widely accessible tomorrow -- just look at the GPS system, or drones. Machine learning has made it so that marketing automation platforms can be predictive -- able to learn, think and act without explicit instructions.

For a decade, one-to-one sales and marketing has lingered in twilight sleep. Still, marketers and salespeople see it coming. Soon, they realize, data intelligence will fill in the details. We'll know the right companies to target, when to do so, what messaging to use, and with whom to share it. The image may be out of focus, but the ingredients are in place.

See all the latest robotics news on Robohub, or sign up for our weekly newsletter.

How well can future health events of patients be predicted from EHR data, at various lengths of time in advance? And how can such predictions improve human health? This talk answers the first question via an approach called high-throughput machine learning, and it speculates about answers to the second question. In particular, this talk argues that many healthcare applications require not just accurate prediction, but accurate prediction by causally-faithful models. Causal discovery from observational data is already a major research direction in machine learning and statistics, and this talk discusses new approaches across the spectrum from when "we know all the relevant variables" to when "we know only one relevant variable" for the task at hand.

At IBM Edge 2016, a team of developers and data scientists presented a practical study that evaluated the efficiency of training a TensorFlow model in a distributed mode. A use case featured high-resolution images of lymph nodes used for possible cancer detection. Relying on a distributed model of TensorFlow and high-performing nature of the OpenPOWER infrastructure, the demonstrated system can accelerate medical data analysis--depending on the number of GPUs and nodes in its cluster. The particular subjects of the research were how training time decreases when the cluster grows and whether the accuracy of the results is affected by the distributed nature of the computations.