Most car company executives appear at the annual Los Angeles Auto Show to tout their brand's new vehicles or crow about its balance sheet. Ford Chief Executive Mark Fields came to town this year to give the kick-off keynote speech at AutoMobility L.A., the industry-only portion of the massive car show, and didn't discuss those things at all. Instead, he talked about the future of mobility, the "city of tomorrow" and how Ford is trying to find its place in a changing transportation landscape that -- confronting autonomous driving, connected cars and ride-sharing -- may soon have no place for traditional car companies. The following is an edited version of the conversation. Ford produces the F-150 truck, the most popular vehicle sold in America, but is also offering a hydrogen fuel cell Clarity and a line of plug-in hybrids.

I have been giving two talks recently on the machine learning workflow, discussing pain points within it and how we might address them. First one was at Spark Summit Europe at Brussels, the other one at MLConf at San Francisco. You can find videos and slides for each below. Main message is that the machine learning workflow is not that simple. That was a great event.

It seems like strange advice: if your phone is full up with rubbish and you want to free some of the storage up, then download even more rubbish. A new trick lets people get extra space on their iPhone just by downloading some new apps, and by tricking the phone into believing that it needs to clear up some space. To do it, all you need to do is to download a big app that takes up more space on your phone than you have left. As soon as you do so, your iPhone will automatically start clearing up much-needed storage space. A man looks at an exhibit entitled'Mimus' a giant industrial robot which has been reprogrammed to interact with humans during a photocall at the new Design Museum in South Kensington, London Electrification Guru Dr. Wolfgang Ziebart talks about the electric Jaguar I-PACE concept SUV before it was unveiled before the Los Angeles Auto Show in Los Angeles, California, U.S The Jaguar I-PACE Concept car is the start of a new era for Jaguar.

Machine Learning is a subfield of computer science by which computers have the ability to learn without being explicitly programmed. Machine learning is a method used to develop complex models and algorithms that lend themselves to prediction. Machine learning is deployed where the system deals with large scale of data. Deployments of machine learning leads to improved speed and accuracy of the functions performed by the system. Machine learning is majorly deployed for solving classification and regression problems.

Anyone who's been licked by a cat can speak of the sandpaper-like surface of their flexible tongues. But according to new research, the tiny hooks responsible for this roughness are the secret behind cats' expert grooming, acting like'heat-seeking missiles' for tangles. By capturing these features with macro and high-speed videography, researchers have now 3D-printed a cat tongue at 400% scale, and they say the technique could be used to improve soft robotics systems and wound cleaning materials. By capturing macro and high speed videos of cats grooming, researchers found that the features on their tongues are like tiny claws in both shape and sharpness. These, they say, act like heat-seeking missiles for tangles.

Your dog remembers more than you might think. A new study that tested the memory of man's best friend found that dogs exhibit something akin to episodic memory -- a process that's been well documented in humans, but difficult to prove in other animals. In experiments, the dogs were able to recall human actions even when they weren't expecting to be tested on what they observed, according to a report published Wednesday in the journal Current Biology. The findings show that episodic memory, thought to be linked to self-awareness, may extend well beyond humans to species outside of the primate lineage. Scientists have long wondered whether other animals have something like episodic memory, which allows us to recall specific past events even though they may not have been particularly important when they happened.

The ongoing evolution in advanced analytics and other emerging technologies is transforming the operational processes in government and public service agencies throughout the world. With the help of these technologies, such organisations are trying to address citizen demands, helping to overcome persistent challenges such as regulatory compliance, outdated legacy IT infrastructures and organisational cultures, according to a new study report from Accenture. In its latest report titled, Emerging Technologies in Public Service, Accenture surveyed nearly 800 public service technology professionals across nine countries to identify emerging technologies being implemented or piloted. The technologies include advanced analytics/predictive modelling, the IoT, intelligent process automation, video analytics, biometrics/identity analytics, machine learning, and natural language processing/generation. The survey found that while more than two-thirds, (70%) of public sector agencies are evaluating the potential of emerging technologies, only a small percentage (25%) is moving beyond the pilot phase to full implementation.

We survey results on neural network expressivity described in "On the Expressive Power of Deep Neural Networks". The paper motivates and develops three natural measures of expressiveness, which all display an exponential dependence on the depth of the network. In fact, all of these measures are related to a fourth quantity, trajectory length. This quantity grows exponentially in the depth of the network, and is responsible for the depth sensitivity observed. These results translate to consequences for networks during and after training. They suggest that parameters earlier in a network have greater influence on its expressive power -- in particular, given a layer, its influence on expressivity is determined by the remaining depth of the network after that layer. This is verified with experiments on MNIST and CIFAR-10. We also explore the effect of training on the input-output map, and find that it trades off between the stability and expressivity.

Since about 100 years ago, to learn the intrinsic structure of data, many representation learning approaches have been proposed, including both linear ones and nonlinear ones, supervised ones and unsupervised ones. Particularly, deep architectures are widely applied for representation learning in recent years, and have delivered top results in many tasks, such as image classification, object detection and speech recognition. In this paper, we review the development of data representation learning methods. Specifically, we investigate both traditional feature learning algorithms and state-of-the-art deep learning models. The history of data representation learning is introduced, while available resources (e.g. online course, tutorial and book information) and toolboxes are provided. Finally, we conclude this paper with remarks and some interesting research directions on data representation learning.

Complex nonlinear models such as deep neural network (DNNs) have become an important tool for image classification, speech recognition, natural language processing, and many other fields of application. These models however lack transparency due to their complex nonlinear structure and to the complex data distributions to which they typically apply. As a result, it is difficult to fully characterize what makes these models reach a particular decision for a given input. This lack of transparency can be a drawback, especially in the context of sensitive applications such as medical analysis or security. In this short paper, we summarize a recent technique introduced by Bach et al. [1] that explains predictions by decomposing the classification decision of DNN models in terms of input variables.