If you are a newcomer to the Deep Learning area, the first question you may have is "Which paper should I start reading from?" Here is a reading roadmap of Deep Learning papers! You will find many papers that are quite new but really worth reading. I would continue adding papers to this roadmap. Editor: What follows is a portion of the papers from this list.

Deep learning is one of the buzziest buzzwords of 2017, and for good reason. Deep learning (more accurately called deep neural networks) attempts to mimic the activities of the brain. The basic principles of neural networks have existed since the late 1950s, yet it wasn't until around 2010 that computers became powerful enough (and data got big enough) for highly complex "deep" neural networks to become practical for real-world applications. Today, this technique is revolutionizing natural language processing and malware detection. Deep learning can figure out how to solve tough problems, such as identifying suspicious online behavior.

FX-7TM offering leverages the company's 7nm FinFET process to deliver best in class IP and Solutions Santa Clara, Calif., June 13, 2017 – GLOBALFOUNDRIES today announced the availability of FX-7 TM, an application-specific integrated circuit (ASIC) offering built on the company's 7nm FinFET process technology. FX-7 is an integrated design platform that combines leading-edge manufacturing process technology with a differentiated suite of intellectual property and 2.5D/3D packaging to deliver the industry's most complete solution for data center, machine learning, automotive, wired communications, and 5G wireless applications. Building on the continued success of FX-14, with industry-leading 56G SerDes and a legacy of ASIC expertise, FX-7 provides a comprehensive suite of tailored interface IPs including High Speed SerDes (60G, 112G), differentiated memory solutions including low-voltage SRAM, high-performance embedded TCAM, integrated DACs/ADCs, ARM processors, and advanced packaging options such as 2.5D/3D. In addition, the FX-7 portfolio enables new design methodologies and complex ASIC solutions for lower power and high-performance applications targeting hyper-scale data centers, 5G networking, and machine and deep learning applications. Future extensions are planned to support solutions for automotive ADAS and imaging applications.

Google's latest effort to thaw relations with China involves an artificial intelligence pow-wow -- and a few games of Go. Google's latest effort to thaw relations with China involves an artificial intelligence pow-wow -- and a few games of Go. Years after Beijing locked out virtually every Alphabet Inc. service, executive chairman Eric Schmidt and a cadre of mid-level Chinese government officials kicked off a summit in the historic canal-laced town of Wuzhen Tuesday: a rare instance of the search leader working in tandem with the country's bureaucrats at a high-profile public event. Google experts and prominent local academics will exchange notes and host discussions but the centerpiece will be a contest between DeepMind's so-far undefeated AlphaGo system and Ke Jie, local champion of the 2,500-year-old strategy board game Go.

IT decision makers are divided about the impact of disruptive technologies such as Artificial Intelligence (AI) and automation - the so called'Fourth Industrial Revolution' - on the labour market, according to new research from BT. Contrary to many reports which speculate about widespread job losses, one third of organisations surveyed who plan to implement AI and automation within the next two years believe it will create more jobs within the workplace. This reflects the view that AI will generate new opportunities for programmers, algorithm designers and software engineers and create new job categories such as AI trainers, ethicists and lawyers. However, the same proportion predict that these technologies could result in job losses in their organisation, given concerns that innovations in robotics and intelligent computer systems may eventually replace jobs traditionally done by humans, particularly those of a manual, repetitive nature. Against this uncertainty surrounding the impact of these technologies on the jobs market, the survey of 1,501 IT decision makers across UK organisations of all sizes revealed that AI and automation is already being implemented by over a third of all respondents. For example, one in four organisations are using automation technologies like drones, robots or autonomous vehicles, with almost two thirds (63%) describing AI or automation technologies as being'very beneficial' to their organisations.

This, then, is a deep neural network: it takes an image input, returns an activity output, and – just as we might learn to detect patterns in puppy behavior without knowing anything about dogs (after seeing enough corgis, we discover common characteristics like fluffy butts and drumstick legs; next, we learn advanced features like splooting) – in between it learns to represent images through hidden layers of representations. Instead of simply taking an image and returning an activity, an RNN also maintains internal memories about the world (weights assigned to different pieces of information) to help perform its classifications. Note that the hidden state computed at time \(t\) (\(h_t\), our internal knowledge) is fed back at the next time step. So what we'd like is for the network to learn how to update its beliefs (scenes without Bob shouldn't change Bob-related information, scenes with Alice should focus on gathering details about her), in a way that its knowledge of the world evolves more gently.

In recent years, we have become increasingly good at training deep neural networks to learn a very accurate mapping from inputs to outputs, whether they are images, sentences, label predictions, etc. from large amounts of labeled data. What our models still frightfully lack is the ability to generalize to conditions that are different from the ones encountered during training. Every time you apply your model not to a carefully constructed dataset but to the real world. The real world is messy and contains an infinite number of novel scenarios, many of which your model has not encountered during training and for which it is in turn ill-prepared to make predictions. The ability to transfer knowledge to new conditions is generally known as transfer learning and is what we will discuss in the rest of this post. Over the course of this blog post, I will first contrast transfer learning with machine learning's most pervasive and successful paradigm, supervised learning. I will then outline reasons why transfer learning warrants our attention. Subsequently, I will give a more technical definition and detail different transfer learning scenarios. I will then provide examples of applications of transfer learning before delving into practical methods that can be used to transfer knowledge.

Attention & interpretability: modern neural networks are hard to interpret because we haven't built tools to make it easy to analyze their decision-making processes. Part of the reason why we haven't built the tools is that it's not entirely obvious how you get a big stack of perceptual math machinery to tell you about what it is thinking in a way that is remotely useful to the untrained eye. The best thing we've been able to come up with, in the case of certain vision and language tasks, is attention where we visualize what parts of a neural network – sometimes down to an individual cell or'neuron' within it – is activating in response to. This can help us diagnose why an AI tool is responding in the way it is. This component is general, working across different neural network architectures (a first, the researchers claim), and only requires the person to fiddle with it at its input or output points.

"65% of a company's business comes from existing customers, and it costs five times as much to attract a new customer than to keep an existing one satisfied." Content & product recommendation systems are important, but they don't go deep enough when it comes to solving CRM problems. To minimize your customer churn rate, you have to dig much deeper into this problem. Recommending the right products to your customers is not nearly enough today to reduce churn rate. We are living in a hyper-connected world.

We present a scalable end-to-end classifier that uses streaming physiological and medication data to accurately predict the onset of sepsis, a life-threatening complication from infections that has high mortality and morbidity. Our proposed framework models the multivariate trajectories of continuous-valued physiological time series using multitask Gaussian processes, seamlessly accounting for the high uncertainty, frequent missingness, and irregular sampling rates typically associated with real clinical data. The Gaussian process is directly connected to a black-box classifier that predicts whether a patient will become septic, chosen in our case to be a recurrent neural network to account for the extreme variability in the length of patient encounters. We show how to scale the computations associated with the Gaussian process in a manner so that the entire system can be discriminatively trained end-to-end using backpropagation. In a large cohort of heterogeneous inpatient encounters at our university health system we find that it outperforms several baselines at predicting sepsis, and yields 19.4% and 55.5% improved areas under the Receiver Operating Characteristic and Precision Recall curves as compared to the NEWS score currently used by our hospital.