Identification of a groundwater contaminant source simultaneously with the hydraulic conductivity in highly-heterogeneous media often results in a high-dimensional inverse problem. In this study, a deep autoregressive neural network-based surrogate method is developed for the forward model to allow us to solve efficiently such high-dimensional inverse problems. The surrogate is trained using limited evaluations of the forward model. Since the relationship between the time-varying inputs and outputs of the forward transport model is complex, we propose an autoregressive strategy, which treats the output at the previous time step as input to the network for predicting the output at the current time step. We employ a dense convolutional encoder-decoder network architecture in which the high-dimensional input and output fields of the model are treated as images to leverage the robust capability of convolutional networks in image-like data processing. An iterative local updating ensemble smoother (ILUES) algorithm is used as the inversion framework. The proposed method is evaluated using a synthetic contaminant source identification problem with 686 uncertain input parameters. Results indicate that, with relatively limited training data, the deep autoregressive neural network consisting of 27 convolutional layers is capable of providing an accurate approximation for the high-dimensional model input-output relationship. The autoregressive strategy substantially improves the network's accuracy and computational efficiency. The application of the surrogate-based ILUES in solving the inverse problem shows that it can achieve accurate inversion results and predictive uncertainty estimates.

Machine learning (ML), artificial intelligence (AI) and other modern statistical methods are providing new opportunities to operationalize previously untapped and rapidly growing sources of data for patient benefit. Whilst there is a lot of promising research currently being undertaken, the literature as a whole lacks: transparency; clear reporting to facilitate replicability; exploration for potential ethical concerns; and, clear demonstrations of effectiveness. There are many reasons for why these issues exist, but one of the most important that we provide a preliminary solution for here is the current lack of ML/AI- specific best practice guidance. Although there is no consensus on what best practice looks in this field, we believe that interdisciplinary groups pursuing research and impact projects in the ML/AI for health domain would benefit from answering a series of questions based on the important issues that exist when undertaking work of this nature. Here we present 20 questions that span the entire project life cycle, from inception, data analysis, and model evaluation, to implementation, as a means to facilitate project planning and post-hoc (structured) independent evaluation. By beginning to answer these questions in different settings, we can start to understand what constitutes a good answer, and we expect that the resulting discussion will be central to developing an international consensus framework for transparent, replicable, ethical and effective research in artificial intelligence (AI-TREE) for health.

Background: Pharmacokinetic evaluation is one of the key processes in drug discovery and development. However, current absorption, distribution, metabolism, excretion prediction models still have limited accuracy. Aim: This study aims to construct an integrated transfer learning and multitask learning approach for developing quantitative structure-activity relationship models to predict four human pharmacokinetic parameters. Methods: A pharmacokinetic dataset included 1104 U.S. FDA approved small molecule drugs. The dataset included four human pharmacokinetic parameter subsets (oral bioavailability, plasma protein binding rate, apparent volume of distribution at steady-state and elimination half-life). The pre-trained model was trained on over 30 million bioactivity data. An integrated transfer learning and multitask learning approach was established to enhance the model generalization. Results: The pharmacokinetic dataset was split into three parts (60:20:20) for training, validation and test by the improved Maximum Dissimilarity algorithm with the representative initial set selection algorithm and the weighted distance function. The multitask learning techniques enhanced the model predictive ability. The integrated transfer learning and multitask learning model demonstrated the best accuracies, because deep neural networks have the general feature extraction ability, transfer learning and multitask learning improved the model generalization. Conclusions: The integrated transfer learning and multitask learning approach with the improved dataset splitting algorithm was firstly introduced to predict the pharmacokinetic parameters. This method can be further employed in drug discovery and development.

Artificial Intelligence (AI) technologies could be broadly categorised into Analytics and Autonomy. Analytics focuses on algorithms offering perception, comprehension, and projection of knowledge gleaned from sensorial data. Autonomy revolves around decision making, and influencing and shaping the environment through action production. A smart autonomous system (SAS) combines analytics and autonomy to understand, learn, decide and act autonomously. To be useful, SAS must be trusted and that requires testing. Lifelong learning of a SAS compounds the testing process. In the remote chance that it is possible to fully test and certify the system pre-release, which is theoretically an undecidable problem, it is near impossible to predict the future behaviours that these systems, alone or collectively, will exhibit. While it may be feasible to severely restrict such systems\textquoteright \ learning abilities to limit the potential unpredictability of their behaviours, an undesirable consequence may be severely limiting their utility. In this paper, we propose the architecture for a watchdog AI (WAI) agent dedicated to lifelong functional testing of SAS. We further propose system specifications including a level of abstraction whereby humans shepherd a swarm of WAI agents to oversee an ecosystem made of humans and SAS. The discussion extends to the challenges, pros, and cons of the proposed concept.

The 2016 U.S. presidential election will go down as one of the most unexpected occurrences in the country's history. According to the Pew Research Center, 73 percent of all voters said that they were shocked by Donald Trump's victory. Even 60 percent of Trump voters didn't expect their candidate to win. The campaign was certainly an interesting moment in U.S. politics, but data visualization firm Periscopic decided to take this election as an opportunity to learn more about emotional expression. What they found was that President Trump expressed more negative facial emotions during his inaugural address than any other president in nearly 40 years.

Today's AI cannot replace humans in cybersecurity but shows promise for driving efficiency and addressing talent shortage, a new report by ProtectWise has shown. Conducted by Osterman Research, the study explores usage trends and sentiments toward AI among more than 400 U.S. security analysts in organizations with 1000 or more employees. Nearly three quarters of respondents have already implemented at least one product that uses AI, but findings uncovered mixed results and a learning curve that needs to be addressed in order to use AI at higher levels of sophistication and effectiveness. " A lot of hype and confusion exists around AI and its role in the cybersecurity industry," said Gene Stevens, CTO, ProtectWise. "In its current state, AI is a tool for driving efficiencies and addressing staffing needs, but it is not going to replace human intelligence any time soon. AI is well positioned today to create machine-accelerated humans: an army of hunters and responders who use a wide array of expert systems to help unearth and prioritize critical threats. In the future, AI will only become more valuable as the industry develops products that improve ease of use and capitalize on AI's efficiency differentiators ."

The National Police Agency on Thursday unveiled a draft bill that would allow vehicles with a high level of autonomous features to run on public roads, with an eye toward implementing the legislation in the first half of 2020. The bill to revise the nation's road traffic law would enable travel for what the government classifies as level 3 autonomous vehicles. Such vehicles can allow drivers to shift their attention elsewhere and let the system drive, except for during emergencies and system glitches that would require them to take back control. In the initial stage, the government may only allow the use of level 3 self-driving technology during highway traffic jams. Autonomous driving technology is classified into five categories.

Evelyn Berezin, who has died aged 93, invented the Data Secretary, the first electronic word processor for secretarial use, and in 1969 founded a company in Hauppauge, Long Island, to manufacture and sell it. She had bumped into the glass ceiling and it was the only way she could get a senior position running a company. The choice of product was tactical. As one of the few women developing computer hardware at the time, she was a two-finger typist and said she had to stay as far away as possible from looking like a secretary. However, she needed something that a small team could create at a price low enough to sell. In the 1960s, most computers were so expensive that companies rented them.

Authorities in the US state of Pennsylvania have given Uber the green light to resume testing self-driving cars, the ride-sharing giant said Tuesday, after a fatal crash in Arizona prompted a pause. Uber said it had received authorization to put autonomous cars back on the road in Pittsburgh, where it has a lab devoted to the technology, but has yet to actually do so. The San Francisco-based company suspended use of self-driving cars in March after one struck and killed a pedestrian in Tempe, Arizona. Authorities in Pennsylvania have given Uber the green light to resume testing self-driving cars, the ride-sharing giant said, after a fatal crash in Arizona prompted a pause. A fatal accident involving a pedestrian and one of Uber's self-driving vehicles could have been prevented, a new report claims.

In the early 1980s, a cluster of fledging computer companies opened up shop in a chaotic corner of northwest Beijing, near the campuses of Peking and Tsinghua Universities. Electronics Street, as the area became known, was a tangle of sturdy bicycles and hand-drawn signs, loud with heated bouts of haggling. Dusty banners hung over pedestrians' heads, while boxes of copy paper stacked 10 or 12 high blocked their path. Pirated software was so abundant that some preferred the moniker Crook Street. The existence of a burgeoning PC market was remarkable, given that many Chinese still did not own a refrigerator. But more remarkable was that the businesses of Electronics Street were private enterprises. Their foray into capitalism was an experiment launched with China's economic reforms, which early on were linked to investments in science and technology.