This is a guest post. The views expressed here are solely those of the author and do not represent positions of IEEE Spectrum or the IEEE. When it comes to new technologies and products, we tend to think of "digital" as synonymous with advanced, modern, and high-def, while "analog" is considered retrograde, outmoded, and low-resolution. But if you think analog is dead, you'd be wrong. Analog processing not only remains at the heart of many vital systems we depend on today, it is now going to make its way into a new breed of compute and intelligent systems that will power some of the most exciting technologies of the future: artificial intelligence and robotics.

Computing has evolved through a series of distinct architectural eras: mainframes, distributed computing, client-server, Internet, and cloud. Each offered a new way of organizing information and connecting people. However, the volume (and velocity and variety) of information is exploding. Today a PC generates only about 90 megabytes of network data per day, but in the not too distant future, a connected autonomous vehicle will produce many terabytes and a connected automated factory will generate over a petabyte. We need more than simply new ways to organize and connect to information, we need new ways to uncover the hidden insights within and to harness the full potential of machines.

LONDON Jaguar Land Rover, Ford and Tata Motors are testing connected cars which can communicate with each other using technology designed to speed up journeys and cut accidents, the first such trials in Britain.

Recent work on discriminative segmental models has shown that they can achieve competitive speech recognition performance, using features based on deep neural frame classifiers. However, segmental models can be more challenging to train than standard frame-based approaches. While some segmental models have been successfully trained end to end, there is a lack of understanding of their training under different settings and with different losses. We investigate a model class based on recent successful approaches, consisting of a linear model that combines segmental features based on an LSTM frame classifier. Similarly to hybrid HMM-neural network models, segmental models of this class can be trained in two stages (frame classifier training followed by linear segmental model weight training), end to end (joint training of both frame classifier and linear weights), or with end-to-end fine-tuning after two-stage training. We study segmental models trained end to end with hinge loss, log loss, latent hinge loss, and marginal log loss. We consider several losses for the case where training alignments are available as well as where they are not. We find that in general, marginal log loss provides the most consistent strong performance without requiring ground-truth alignments. We also find that training with dropout is very important in obtaining good performance with end-to-end training. Finally, the best results are typically obtained by a combination of two-stage training and fine-tuning.

In a physical neural system, where storage and processing are intimately intertwined, the rules for adjusting the synaptic weights can only depend on variables that are available locally, such as the activity of the pre- and post-synaptic neurons, resulting in local learning rules. A systematic framework for studying the space of local learning rules is obtained by first specifying the nature of the local variables, and then the functional form that ties them together into each learning rule. Such a framework enables also the systematic discovery of new learning rules and exploration of relationships between learning rules and group symmetries. We study polynomial local learning rules stratified by their degree and analyze their behavior and capabilities in both linear and non-linear units and networks. Stacking local learning rules in deep feedforward networks leads to deep local learning. While deep local learning can learn interesting representations, it cannot learn complex input-output functions, even when targets are available for the top layer. Learning complex input-output functions requires local deep learning where target information is communicated to the deep layers through a backward learning channel. The nature of the communicated information about the targets and the structure of the learning channel partition the space of learning algorithms. We estimate the learning channel capacity associated with several algorithms and show that backpropagation outperforms them by simultaneously maximizing the information rate and minimizing the computational cost, even in recurrent networks. The theory clarifies the concept of Hebbian learning, establishes the power and limitations of local learning rules, introduces the learning channel which enables a formal analysis of the optimality of backpropagation, and explains the sparsity of the space of learning rules discovered so far.

If there is anything that chip giant Intel has learned over the past two decades as it has gradually climbed to dominance in processing in the datacenter, it is ironically that one size most definitely does not fit all. As the tight co-design of hardware and software continues in all parts of the IT industry, we can expect fine-grained customization for very precise – and lucrative – workloads, like data analytics and machine learning, just to name two of the hottest areas today. Software will run most efficiently on hardware that is tuned for it, although we are used to thinking of that process in a mirror image, where programmers tweak their code to take advantage of the forward-looking features a chip maker conceives of four or five years before they are etched into its transistors and delivered as a product. The competition is fierce these days, and Intel has to move fast if it is to keep its compute hegemony in the datacenter. That is why at the Intel Developer Forum in San Francisco the company put a new path on the Knights family of many-core processors that will see the company deliver a version of this chip specifically tuned for machine learning workloads.

In the past half decade, artificial intelligence and machine learning have made significant leaps into the mainstream and into our daily lives. According to research firm Markets and Markets, the artificial intelligence market is set to grow to 5.05 billion by 2020 thanks to the increased applicability of various AI technologies into everything from finance to healthcare to retail. Today, doctors can diagnose Sepsis with an AI algorithm, for instance, and researchers can track endangered species through AI-enhanced photo capture systems. Clearly, these new self-learning and ever-improving technologies have limitless potential in a number of innovative industries. The U.S. Chamber of Commerce's Technology Engagement Center (C_TEC) recently hosted a panel discussion during its TecNation 2016 event that focused on where we stand with Artificial Intelligence and how it will affect our lives and unlock our potential in the long run.

Israeli startup Faception made headlines this year by claiming it could predict how likely people are to be terrorists, pedophiles, and more by analyzing faces with deep learning. Experts and research in the field, however, suggest that it is more fantasy than reality. Faception assigns ratings after training artificial intelligence on faces of terrorists, pedophiles, Mensa members, professional poker players, and more. Through deep learning--that emerging technique found in everything from Alpha Go to Siri to Netflix--the AI can supposedly predict how likely a new face is to belong to any given group. While this may sound believable, there's no evidence that face-based personality predictions are more than a tiny bit accurate.

Deep learning is currently one of the best providers of solutions regarding problems in image recognition, speech recognition, object recognition, and natural language with its increasing number of libraries that are available in Python. The aim of deep learning is to develop deep neural networks by increasing and improving the number of training layers for each network, so that a machine learns more about the data until it's as accurate as possible. Developers can avail the techniques provided by deep learning to accomplish complex machine learning tasks, and train AI networks to develop deep levels of perceptual recognition. Deep learning is the next step to machine learning with a more advanced implementation. Currently, it's not established as an industry standard, but is heading in that direction and brings a strong promise of being a game changer when dealing with raw unstructured data.

RTB House, a technology company specializing in retargeting scenarios, has come up with a brand new model that relies on deep learning (currently the most promising subfield of AI-oriented research) to craft digital features that recognize the attitude, intention and intent of internet users. It allows for accurate estimation of the conversion probability, which in turn makes personalized retargeting more efficient than before. The model can even be applied to users who haven't clicked ads, a long-sought after feature of digital marketers. Users take hundreds of small steps when visiting advertiser's website. The model developed by RTB House uses deep learning to identify every one of these footprints, in order to find patterns in decision-making.