We revisit the importance of the individual units in Convolutional Neural Networks (CNNs) for visual recognition. By conducting unit ablation experiments on CNNs trained on large scale image datasets, we demonstrate that, though ablating any individual unit does not hurt overall classification accuracy, it does lead to significant damage on the accuracy of specific classes. This result shows that an individual unit is specialized to encode information relevant to a subset of classes. We compute the correlation between the accuracy drop under unit ablation and various attributes of an individual unit such as class selectivity and weight L1 norm. We confirm that unit attributes such as class selectivity are a poor predictor for impact on overall accuracy as found previously in recent work \cite{morcos2018importance}. However, our results show that class selectivity along with other attributes are good predictors of the importance of one unit to individual classes. We evaluate the impact of random rotation, batch normalization, and dropout to the importance of units to specific classes. Our results show that units with high selectivity play an important role in network classification power at the individual class level. Understanding and interpreting the behavior of these units is necessary and meaningful.

In a paper from California Institute of Technology's Jet Propulsion Laboratory, the author laments the loss of connection between machine learning and solving real-world problems. The main question that seems to come up is: what good can machine learning or artificial intelligence (AI) do? A lot, as it turns out. AI could play a big role in aiding and even solving the issues of world hunger and poverty in ways that are surprisingly simple, but increasingly necessary. When you think of AI, it's possible that fixing world hunger isn't the first application that comes to mind.

Mapping a city's buildings might seem like a simple task, one that could be easily automated by training a computer to read satellite photos. Because buildings are physically obvious facts out in the open that do not move around, they can be recorded by the satellites circling our planet. Computers can then "read" these satellite photographs, which are pixelated images like everyday photographs except that they carry more information about the light waves being reflected from various surfaces. That information can help determine the kind of building material and even plant species that appears in an image. Other patterns match up with predictable objects, like the straight lines of roads or the bends of rivers.

A slew of AI weed killers are on the horizon and have the potential to disrupt the multibillion dollar pesticides business. Among them is Swiss-company ecoRobotix and its weed-killing robot. It's solar-powered and can kill weeds for 12 hours straight without an operator at the helm. EcoRobotix uses 20 times less herbicide than traditional methods that spray entire fields. Founded in 2011, ecoRobotix develops autonomous weeding robots, which help farmers to produce healthier food with a more efficient and sustainable use of herbicides.

When the 300-foot Maersk Launcher docked in San Diego early Monday morning, it unloaded a cargo of hardened black blobs scooped from the bottom of the sea. The blobs are not rocks, but naturally-occurring metallic nodules that could one day yield metal deposits of cobalt, manganese, and nickel--not to mention scarce rare earth minerals. As worldwide demand rises for electric vehicle batteries and wind turbines, along with next generation technologies and weapon systems, demand for these metals has taken off. And the seabed is a prime target for those mining operations. Of course, it's no small feat to bring these potato-sized nodules from the bottom of the remote Pacific Ocean, and then sail them to a processing plant where the metals can be extracted.

UK Prime Minister Theresa May has said she is disappointed by the US's "unjustified decision" to slap tariffs on EU steel and aluminium. The tariffs of 25% on steel and 10% on aluminium, which affect the EU, Canada and Mexico, came into effect on Friday. All three are planning retaliatory moves. Mrs May said the EU and UK should be exempted and would work together to "protect and safeguard our workers and industries". UK Steel said the tariffs, which apply to a wide range of steel and aluminium products such as sheets, plates, bars, pipes and "semi-finished" products, will damage not only the UK steel sector but also the US economy.

A new robot created in Nova Scotia may mean farmers could get some help tackling troublesome weeds in their fields. This month, Nexus Robotics, a technology startup based in Dartmouth, N.S., won the weed-and-feed competition at the agBOT Challenge, an international showdown between agricultural robots in Rockville, Ind. Dubbed R2 Weed2 or Hal-Bot, the autonomous machine uses artificial intelligence to distinguish between weeds and crops and is designed to both pluck weeds and spray herbicide. "We want to get rid of the weed and keep the crop and even fertilize it. So one of the advancements … we made is vision systems can be better than humans at distinguishing them," said Thomas Trappenberg, part of the small team behind the battery-powered robot. VIDEO: Halifax startup takes on big agriculture corporations, and wins international robotics competition.

The demand of metals by modern technology has been shifting from common base metals to a variety of minor metals, such as cobalt or indium. The industrial importance and limited geological availability of some minor metals have led to them being considered more "critical," and there is a growing interest in such critical metals and their producing companies. In this research, we create a novel framework, Dynamic Advisor-Based Ensemble (dynABE), to predict the stock trend of major critical metal producers. Specifically, dynABE first utilizes domain knowledge to group the features into different "advisors," each advisor dealing with a particular economic sector. Then through ensembles of weak classifiers, each advisor produces a prediction result, and all the advisors are combined again in a biased online update fashion to dynamically make the final prediction. Based on a misclassification error of 32% for Jinchuan Group's stock (HKG: 2362), we further test a simple stock trading strategy, which leads to a back-tested return of 296%, or an excess return of 130% within one year. In addition, the feature set selected by dynABE also suggests potentially influential factors to metal criticality, because stock prices of major producers influence metal production. Therefore, not only does this research propose a novel framework for specialized stock trend prediction, it also provides domain insights into dynamic features that potentially influence metal criticality.

Performance monitoring, anomaly detection, and root-cause analysis in complex cyber-physical systems (CPSs) are often highly intractable due to widely diverse operational modes, disparate data types, and complex fault propagation mechanisms. This paper presents a new data-driven framework for root-cause analysis, based on a spatiotemporal graphical modeling approach built on the concept of symbolic dynamics for discovering and representing causal interactions among sub-systems of complex CPSs. We formulate the root-cause analysis problem as a minimization problem via the proposed inference based metric and present two approximate approaches for root-cause analysis, namely the sequential state switching ($S^3$, based on free energy concept of a restricted Boltzmann machine, RBM) and artificial anomaly association ($A^3$, a classification framework using deep neural networks, DNN). Synthetic data from cases with failed pattern(s) and anomalous node(s) are simulated to validate the proposed approaches. Real dataset based on Tennessee Eastman process (TEP) is also used for comparison with other approaches. The results show that: (1) $S^3$ and $A^3$ approaches can obtain high accuracy in root-cause analysis under both pattern-based and node-based fault scenarios, in addition to successfully handling multiple nominal operating modes, (2) the proposed tool-chain is shown to be scalable while maintaining high accuracy, and (3) the proposed framework is robust and adaptive in different fault conditions and performs better in comparison with the state-of-the-art methods.

Autonomous luggage maker ForwardX Robotics today announced it has raised $10 million to bring its suitcase Ovis to market. At $399, the luggage can move a maximum 6.2 miles per hour and will ship to its first customers in late 2018. ForwardX was founded in 2016, but its luggage initially grabbed the world's attention in January at the Consumer Electronics Show (CES) 2018 in Las Vegas. The 9.9 lb suitcase is made of polypropylene and carbon fiber and is able to follow you by deploying computer vision that tracks your body and face, even if you are momentarily out of sight. Though Ovis has been tested and found to be useful in environments outside airports, like city streets, its battery only lasts for four hours of use, and it must be switched to the old-fashioned manual mode on escalators since it cannot yet handle moving stairs.