IBM's Watson supercomputer has beat Jeopardy champions, reconstituted recipes, and even helped create highlight reels for the World Cup. Now it's taking on a new tech challenge; changing how the construction industry operates. A new partnership between IBM and Fluor, a global engineering and construction company, will put the supercomputer's computational skills to work on making building more efficient. The new Watson-based system, in development since 2015 and now in use on select projects, will be able to analyze a job site "like a doctor diagnoses a patient," according to Leslie Lindgren, Fluor's vice president of Information Management. That degree of risk analysis, predictive logistics, and comprehension is no small challenge given the complexity of today's construction megaprojects.

We propose a new notion of'non-linearity' of a network layer with respect to an input batch that is based on its proximity to a linear system, which is reflected in the nonnegative rank of the activation matrix. Considering batches of similar samples, we find that high non-linearity in deep layers is indicative of memorization. Furthermore, by applying our approach layer-by-layer, we find that the mechanism for memorization consists of distinct phases. We perform experiments on fully-connected and convolutional neural networks trained on several image and audio datasets. Our results demonstrate that as an indicator for memorization, our technique can be used to perform early stopping. A fundamental challenge in machine learning is balancing the bias-variance tradeoff, where overly simple learning models underfit the data (suboptimal performance on the training data) and overly complex models are expected to overfit or memorize the data (perfect training set performance, but suboptimal test set performance). The latter direction of this tradeoff has come into question with the observation that deep neural networks do not memorize their training data despite having sufficient capacity to do so (Zhang et al., 2016), the explanation of which is a matter of much interest. Due to their convenient gradient properties and excellent performance in practice, rectified-linear units (ReLU) have been widely adopted and are now ubiquitous in the field of deep learning.

Abstract--In this work, we present a novel strategy for correcting imperfections in occupancy grid maps called map decay. The objective of map decay is to correct invalid occupancy probabilities of map cells that are unobservable by sensors. The strategy was inspired by an analogy between the memory architecture believed to exist in the human brain and the maps maintained by an autonomous vehicle. It consists in merging sensory information obtained during runtime (online) with a priori data from a high-precision map constructed offline. In map decay, cells observed by sensors are updated using traditional occupancy grid mapping techniques and unobserved cells are adjusted so that their occupancy probabilities tend to the values found in the offline map. This strategy is grounded in the idea that the most precise information available about an unobservable cell is the value found in the high-precision offline map. Map decay was successfully tested and is still in use in the IARA autonomous vehicle from Universidade Federal do Espírito Santo. The brain allows humans to operate in highly dynamic and complex environments, and to solve general purpose problems. The idea of giving these abilities to artificial entities by reproducing the brain's cognitive processes always fascinated researchers.

Many developers, including myself, want to use AI in their applications. Building Machine Learning models, however, often requires a lot of expertise and time. This article describes a technique called AutoML, which can be used by developers to build models without having to be data scientists. While developers only have to provide the data and define the goals, AutoML figures out the best model automatically. Cognitive services are provided by most cloud providers these days.

Marketers get excited about data, artificial intelligence and the internet of things because of their combined power to potentially impact consumers' everyday lives. Across the commerce landscape, the potential applications may be limitless: Farmers are now using satellite data to help increase crop yields and improve the quality of the food we eat. Shippers are deploying blockchain technology to modernize the supply chain and get products into stores more safely and quickly. Banks are relying on encrypted mainframe computers to help protect consumers' personal data and prevent cybercrime. One of the areas in which marketers have only just begun to tap the exponentially increasing unstructured data of the internet is the weather.

One of the most difficult challenges faced by businesses in asset-intensive industries is how to control and scale the half billion and growing "smart" devices that make up the Internet of Things (IoT)? As much as 80 percent of IoT data in any organization is unstructured. And, let's be honest, "smart" devices really aren't that smart yet. As part of its giant rollout of AI solutions pre-trained for specific industries and professions, IBM Services is launching a new Connected Manufacturing offering that includes a method and approach to help clients accelerate their IoT transformation–from strategy, implementation, and security to managed services and ongoing operations. This combined capability, IBM said, will help its clients connect all of their manufacturing equipment, sensors, and systems to enable business improvement across OEE, quality, lead times and productivity.

IBM is introducing a number of "AI-powered solutions" intended help brands better manage the balance between human intelligence and machine learning. For marketers, IBM's WEATHERfx Footfall with Watson represents a new AI ad solution that's specifically designed to help drive in-store traffic by optimizing marketing campaigns. By improving conversion rates with contextually relevant advertising to reach the right audience at the right time, Subway successfully increased sales and attracted new clients. Subway's WEATHERfx Footfall with Watson work involved designing ads based on shifting weather patterns. "Why advertise hot sandwiches when it's 100 degrees out?" is the thinking behind the creative ad units.

Bill Delrieu: In traditional analytics, you have a theory and you go through all your data and workloads to either validate that theory or completely disprove it. Then you come up with an insight and repeat that process over and over again. In the augmented or predictive analytics model, machine learning tools can provide insights without your having to repeat that process over and over again. These insights can be really useful in helping you understand what customers are doing and how to best serve up experiences to them. An example is the Google Analytics Intelligence feature that automatically generates insights.

IBM has ushered its customers into the era of enterprise data science for more than a decade, starting with the investment of the ILOG CPLEX and SPSS acquisitions. As the data science market evolved, new macro trends developed, and IBM invested in advanced technologies and platforms to respond to this shift. In 2016, IBM introduced Data Science Experience and several Watson offerings, which blurred the lines between our new and old technologies. We have now made the decision to simplify our portfolio for our customers under one single brand -- IBM Watson Studio. IBM Watson Studio was first announced in the IBM Public Cloud at our Think Conference in March 2018, which included the integration of the capabilities of Data Science Experience Cloud and a new interface for SPSS Modeler.

The IBM Watson AI XPRIZE is a four-year competition where teams work to improve the world with artificial intelligence. The competition began in 2017 with 148 problem domains in sustainability, artificial general intelligence, education, and a variety of other grand challenge areas. 59 teams advanced to the second year of the competition and ten teams earned special recognition as “milestone nominees.” The properties of the advancing problem domains highlight opportunities and challenges for the “AI for Good” movement. We detail the judging process and highlight preliminary results from cutting the field of competing teams.