We demonstrate a simple strategy to cope with missing data in sequential inputs, addressing the task of multilabel classification of diagnoses given clinical time series. Collected from the pediatric intensive care unit (PICU) at Children's Hospital Los Angeles, our data consists of multivariate time series of observations. The measurements are irregularly spaced, leading to missingness patterns in temporally discretized sequences. While these artifacts are typically handled by imputation, we achieve superior predictive performance by treating the artifacts as features. Unlike linear models, recurrent neural networks can realize this improvement using only simple binary indicators of missingness. For linear models, we show an alternative strategy to capture this signal. Training models on missingness patterns only, we show that for some diseases, what tests are run can be as predictive as the results themselves.

Recent sequential pattern mining methods have used the minimum description length (MDL) principle to define an encoding scheme which describes an algorithm for mining the most compressing patterns in a database. We present a novel subsequence interleaving model based on a probabilistic model of the sequence database, which allows us to search for the most compressing set of patterns without designing a specific encoding scheme. Our proposed algorithm is able to efficiently mine the most relevant sequential patterns and rank them using an associated measure of interestingness. The efficient inference in our model is a direct result of our use of a structural expectation-maximization framework, in which the expectation-step takes the form of a submodular optimization problem subject to a coverage constraint. We show on both synthetic and real world datasets that our model mines a set of sequential patterns with low spuriousness and redundancy, high interpretability and usefulness in real-world applications. Furthermore, we demonstrate that the quality of the patterns from our approach is comparable to, if not better than, existing state of the art sequential pattern mining algorithms.

We develop a supervised machine learning model that detects anomalies in systems in real time. Our model processes unbounded streams of data into time series which then form the basis of a low-latency anomaly detection model. Moreover, we extend our preliminary goal of just anomaly detection to simultaneous anomaly prediction. We approach this very challenging problem by developing a Bayesian Network framework that captures the information about the parameters of the lagged regressors calibrated in the first part of our approach and use this structure to learn local conditional probability distributions.

Investors believe it is'inevitable' that artificial intelligence will destroy millions of jobs and that governments are unprepared for such an impact, according to a new survey. Artificial intelligence (AI), or the process by which computers or robots take on tasks that need human intelligence, is one of the key themes of this week's Web Summit in Lisbon. The poll among 224 venture capitalists attending the conference showed 53 percent believed AI would destroy millions of jobs and 93 percent saw governments as unprepared for this. The poll among 224 venture capitalists attending the Web summit in Lisbon found 53 percent believed AI would destroy millions of jobs and 93 percent saw governments as unprepared for this. The survey also found that 83 percent of the investors canvassed expect Britain's exit from the European Union to damage Europe's economy and 77 percent believe it will damage British startups.

Views from the Marketplace are paid for by advertisers and select partners of MIT Technology Review. Is your organization using, or planning to adopt, machine learning? If so, please share your experiences and insights in this survey. And even if you have no plans to use machine learning, please take the survey anyway--we'd love to know why.

Human resources departments rarely, if ever, are thought of as cutting edge when it comes to the use of technology. A closer look, however, shows the implementation of new technologies, including solutions powered by Artificial Intelligence (AI), in almost every aspect of the talent function. According to a recent Towers Watson HR Service Delivery and Technology Survey, HR professionals are overhauling structure to improve quality and efficiency with 33% of the group spending significantly more on technology in the last year. HR's investment in new technology has also spurred the creation of new data sources. Data around employee productivity, wellness, manager effectiveness, and a host of other activities is quickly dwarfing the traditional data set that HR has traditionally been using.

Wipro Cited as a Leader in Service Providers for Next-generation Oracle Application Projects by Leading Independent Research and Advisory Firm Business Wire India Wipro Ltd. (NYSE: WIT, BSE: 507685, NSE: WIPRO), a leading global information technology, consulting and business process services company today announced that Wipro has been cited as a "Leader" by technology global research and advisory firm Forrester Research Inc. in its report, 'The Forrester Wave(TM): Services Providers for Next-Generation Oracle Application Projects, Q3 2016'. Forrester evaluated 13 service providers across three categories of current offering, strategy and market presence. Forrester wrote, "Wipro is a proven provider making big investments in the shift to cloud. It has a long track record of delivering Oracle services ... [Wipro] was recognized by Oracle with a co-innovation award at Oracle's Collaborate conference in 2016." According to the Forrester report: "Wipro wants to be the partner of choice for tomorrow's digital businesses; it has invested in its own artificial intelligence tool, [Wipro HOLMES Artificial Intelligence Platform(TM)], and in next-generation Oracle technologies. Wipro has recently started to invest in digital experience and customer journey mapping capabilities, including its acquisition of Designit and associated studios."

During a lab meeting, one of our PhD researchers recalls how her father would forbid her from using paper to help solve maths homework problems by writing them down. Another admits that she sometimes still uses her hands to make small calculations, although she does so while hiding them behind her back. When we realise that all of us use our fingers in order to answer demands for the'third, fifth, and seventh digits' of our secret online banking password, we laugh in relief. We are not so daft after all, or at least we are not alone. Consider a game of Scrabble, the researchers say.

RiskIQ, a startup with a new kind of security technology, has raised $30.5 million in a third round of funding. Georgian Partners led the round, with participation from existing investors Summit Partners, Battery Ventures, and MassMutual Ventures. RiskIQ notes that threats outside the firewall are vast and dynamic, so the company provides clients with access to the widest range of security intelligence and applications necessary to understand exposures and how to take action. RiskIQ is one of many companies currently applying machine learning to security. The San Francisco-based company will use this capital to expand its platform, sales, and digital risk applications.

Cape Analytics has raised $14 million to use computer vision and machine learning to improve automated property underwriting for insurance companies. Formation 8 led the round, with participation from XL Innovate, Data Collective, Lux Capital, Khosla Ventures, Promus Ventures, and Montage Ventures. The funding is one more application of machine learning and computer vision for business automation. Palo Alto, Calif.-based Cape Analytics starts with images of a home to help assess the home's value and allow an insurance carrier to deliver more accurate and fast quotes. The funding will allow Cape Analytics to expand its world-class engineering and sales teams as it brings its proprietary data to more regions and customers around the United States.