"To me, it's not just about automating the rig, it's about automating everything upstream of the rig," says Ahmed Hashmi, head of upstream technology for BP Plc." Using deep learning to listen for early warning signs that a car might be nearing a breakdown. "Jeff Dean, who leads the Google Brain research group, mused last week that some of the work of such workers could be supplanted by software. He described what he termed "automated machine learning" as one of the most promising research avenues his team was exploring." Andrew Ng demonstrates Baidu's new office entrance!

Summary: A major problem with chatbots is that they can only provide information from what's in their knowledge base. Here's a new approach that makes your chatbot smarter with every question it can't answer, making it a self-learning lifelong learner. If you've been keeping up with the explosive growth in chatbots you probably already know that there are two basic architectures: They are relatively simple and fast to build, with decision-tree or waterfall-like logic structures of predefined queries and responses. AI Chatbots use deep learning engines to formulate responses. They do not have rigidly defined structures and are able to learn conversational responses after some initial training.

This is the second article in the four-part series on History of Artificial Intelligence. The first part can be accessed here. Every decade seems to have its technological buzzwords: we had personal computers in the 1980s; Internet and worldwide web in 1990s; smartphones and social media in 2000s; and Artificial Intelligence (AI) and Machine Learning in this decade. The 1950-82 era saw a new field of Artificial Intelligence (AI) being born, a lot of pioneering research being done, massive hype being created, and AI going into hibernation when this hype did not materialize, and the research funding dried up [56]. During 1983 and 2010, research funding ebbed and flowed, and research in AI continued to gather steam although " some computer scientists and software engineers would avoid the term artificial intelligence for fear of being viewed as wild-eyed dreamers" [43].

Deep multitask learning boosts performance by sharing learned structure across related tasks. This paper adapts ideas from deep multitask learning to the setting where only a single task is available. The method is formalized as pseudo-task augmentation, in which models are trained with multiple decoders for each task. Pseudo-tasks simulate the effect of training towards closely-related tasks drawn from the same universe. In a suite of experiments, pseudo-task augmentation is shown to improve performance on single-task learning problems. When combined with multitask learning, further improvements are achieved, including state-of-the-art performance on the CelebA dataset, showing that pseudo-task augmentation and multitask learning have complementary value. All in all, pseudo-task augmentation is a broadly applicable and efficient way to boost performance in deep learning systems.

Interpreting black box classifiers, such as deep networks, allows an analyst to validate a classifier before it is deployed in a high-stakes setting. A natural idea is to visualize the deep network's representations, so as to "see what the network sees". In this paper, we demonstrate that standard dimension reduction methods in this setting can yield uninformative or even misleading visualizations. Instead, we present DarkSight, which visually summarizes the predictions of a classifier in a way inspired by notion of dark knowledge. DarkSight embeds the data points into a low-dimensional space such that it is easy to compress the deep classifier into a simpler one, essentially combining model compression and dimension reduction. We compare DarkSight against t-SNE both qualitatively and quantitatively, demonstrating that DarkSight visualizations are more informative. Our method additionally yields a new confidence measure based on dark knowledge by quantifying how unusual a given vector of predictions is.

Deep Q-learning is investigated as an end-to-end solution to estimate the optimal strategies for acting on time series input. Experiments are conducted on two idealized trading games. 1) Univariate: the only input is a wave-like price time series, and 2) Bivariate: the input includes a random stepwise price time series and a noisy signal time series, which is positively correlated with future price changes. The Univariate game tests whether the agent can capture the underlying dynamics, and the Bivariate game tests whether the agent can utilize the hidden relation among the inputs. Stacked Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM) units, Convolutional Neural Network (CNN), and multi-layer perceptron (MLP) are used to model Q values. For both games, all agents successfully find a profitable strategy. The GRU-based agents show best overall performance in the Univariate game, while the MLP-based agents outperform others in the Bivariate game.

Artificial intelligence is revolutionizing our lives at an ever increasing pace. At the heart of this revolution is the recent advancements in deep neural networks (DNN), learning to perform sophisticated, high-level tasks. However, training DNNs requires massive amounts of data and is very computationally intensive. Gaining analytical understanding of the solutions found by DNNs can help us devise more efficient training algorithms, replacing the commonly used mthod of stochastic gradient descent (SGD). We analyze the dynamics of SGD and show that, indeed, direct computation of the solutions is possible in many cases. We show that a high performing setup used in DNNs introduces a separation of time-scales in the training dynamics, allowing SGD to train layers from the lowest (closest to input) to the highest. We then show that for each layer, the distribution of solutions found by SGD can be estimated using a class-based principal component analysis (PCA) of the layer's input. This finding allows us to forgo SGD entirely and directly derive the DNN parameters using this class-based PCA, which can be well estimated using significantly less data than SGD. We implement these results on image datasets MNIST, CIFAR10 and CIFAR100 and find that, in fact, layers derived using our class-based PCA perform comparable or superior to neural networks of the same size and architecture trained using SGD. We also confirm that the class-based PCA often converges using a fraction of the data required for SGD. Thus, using our method training time can be reduced both by requiring less training data than SGD, and by eliminating layers in the costly backpropagation step of the training.

There is new code to play around with for those interested in machine learning and space, a model that predicts hilarious search trends for sex site YouPorn, and another funny story about an ostensibly intelligent medical chatbot in New Zealand. Hunting exoplanets with ML – The machine learning code that a Google engineer and an astrophysicist used to detect exoplanets has been published online. Christopher Shallue, a senior software engineer at Google, and Andrew Vanderburg, a postdoctoral fellow studying astrophysics at the University of Texas, USA, discovered another planet lurking in the Kepler-90 system. It was a special find. Not only was it spotted using a convolutional neural network, but it meant that the Solar System was no longer the biggest planetary system found so far.

Like "big data," today's "artificial intelligence" is all about new tools and approaches for making sense of and profiting from the ever-increasing flood of data. It is the most recent stage in the steady evolution of computer technology since the late 1940s, which has been driven by the increasingly sophisticated and varied use of the key product of computers--digital data. Specifically, today's "artificial intelligence" is the latest stage in the evolution of the fruitful marriage of computer engineering and statistics, of teaching computers to learn from data, or "machine learning." What is frequently overlooked is that this has been--and will continue to be--an evolution in how humans teach machines to learn and how humans learn from automated processes to improve their work. In other words, we are witnessing today a new stage in the steady progress over the last seventy years in the scale and scope of augmented intelligence.

The nation has recently embarked on an all government approach to the problem of cancer, codified in the'Cancer Moonshot' initiative of the Obama administration led by Vice President Biden. Cancer is an extremely complex disease, which disrupts basic biological processes at a fundamental level, leading to renegade cells threatening the health of an individual. To accelerate the capabilities needed to realize the promise envisioned for the Cancer Moonshot and to establish a new paradigm for cancer research for years to come, the Department of Energy (DOE) entered into a partnership with the National Cancer Institute (NCI) of the National Institutes of Health (NIH). This partnership identified three key challenges that the combined resources of DOE and NCI can accelerate: to provide better understanding of the disease, to make effective use of the ever-growing volumes and diversity of cancer related data to build predictive models, and, ultimately, to provide guidance and support decisions on anticipated effective treatments for individual patients. Four DOE national laboratories are collaborating with the NCI and the NCI-supported Frederick National Laboratory for Cancer Research to advance these challenges.