As machine learning models become increasingly larger, trained weakly supervised on large, possibly uncurated data sets, it becomes increasingly important to establish mechanisms for inspecting, interacting, and revising models to mitigate learning shortcuts and guarantee their learned knowledge is aligned with human knowledge. The recently proposed XIL framework was developed for this purpose, and several such methods have been introduced, each with individual motivations and methodological details. In this work, we provide a unification of various XIL methods into a single typology by establishing a common set of basic modules. In doing so, we pave the way for a principled comparison of existing, but, importantly, also future XIL approaches. In addition, we discuss existing and introduce novel measures and benchmarks for evaluating the overall abilities of a XIL method. Given this extensive toolbox, including our typology, measures, and benchmarks, we finally compare several recent XIL methods methodologically and quantitatively. In our evaluations, all methods prove to revise a model successfully. However, we found remarkable differences in individual benchmark tasks, revealing valuable application-relevant aspects for integrating these benchmarks in developing future methods.

Pendulo Studios' Vertigo begins, just like the 1958 film, with a visual and musical motif of spirals. Round and round they go until you meet author Ed Miller in the worst moment of his life. Ed narrowly survives a car crash, but he loses his wife, Faye and their daughter. Staring down at the wreck of his car in a ravine, Ed suffers a debilitating bout of vertigo, only to relive the suicide of his father shortly after. A little later, you step into the shoes of Dr Julia Lomas, a therapist called in to deal with Ed's vertigo and why he keeps talking about a wife and child whom no one but him seems to recall. While it's called Vertigo, complete with the licence of Hitchcock's name and likeness, the game makes hamfisted references to the director's work.

Hate speech detection is a common downstream application of natural language processing (NLP) in the real world. In spite of the increasing accuracy, current data-driven approaches could easily learn biases from the imbalanced data distributions originating from humans. The deployment of biased models could further enhance the existing social biases. But unlike handling tabular data, defining and mitigating biases in text classifiers, which deal with unstructured data, are more challenging. A popular solution for improving machine learning fairness in NLP is to conduct the debiasing process with a list of potentially discriminated words given by human annotators. In addition to suffering from the risks of overlooking the biased terms, exhaustively identifying bias with human annotators are unsustainable since discrimination is variable among different datasets and may evolve over time. To this end, we propose an automatic misuse detector (MiD) relying on an explanation method for detecting potential bias. And built upon that, an end-to-end debiasing framework with the proposed staged correction is designed for text classifiers without any external resources required.

Due to the overwhelming response to our previous expert paper suggestion blog, we had to do another. We asked some of our expert community the papers they would suggest everybody read when working in the field. Haven't seen the first blog? You can read the recommendations of Andrew Ng, Jeff Clune, Myriam Cote and more here. Alexia suggested this paper as it explains how many classifiers can be thought of as estimating an f-divergence.

Artificial Intelligence, also known as AI, is anticipated to become at least a $65 billion market in the next five years. Currently, it is almost impossible to find an established organisation that is not investing in AI technology. However, it seems like enough attention is not provided to this technology before it is availed to the users. There have been major catastrophes encountered due to the failure of AI, and this has made tech leaders legitimately worried about it. This article is little on the negative side, but necessary.

The Internet of Things (IoT) appeared all over the news throughout 2017, whether that was due to the development of new devices, cyber attacks using unsecured devices or even new IoT divisions from companies like Dell and Rolls-Royce, we have definitely heard a lot about the emerging sector. Of course this comes as no surprise, as the oft-quoted Gartner prediction goes: there will be over 20 billion connected things by 2020. The increase in connected devices over the past year, from toasters to toothbrushes, shows we may be getting even closer to the forecast. As the opportunity within the IoT sector continues to rise, businesses globally have been taking a leap to developing unique devices or searching for a way to get in on the action with emerging software or platform solutions. Here are some IoT trends to watch out for in 2018, according to the experts.

The Internet of Things (IoT) appeared all over the news throughout 2017, whether that was due to the development of new devices, cyber attacks using unsecured devices or even new IoT divisions from companies like Dell and Rolls-Royce, we have definitely heard a lot about the emerging sector. Of course this comes as no surprise, as the oft-quoted Gartner prediction goes: there will be over 20 billion connected things by 2020. The increase in connected devices over the past year, from toasters to toothbrushes, shows we may be getting even closer to the forecast. As the opportunity within the IoT sector continues to rise, businesses globally have been taking a leap to developing unique devices or searching for a way to get in on the action with emerging software or platform solutions. Here are some IoT trends to watch out for in 2018, according to the experts.

The Internet of Things (IoT) appeared all over the news throughout 2017, whether that was due to the development of new devices, cyber attacks using unsecured devices or even new IoT divisions from companies like Dell and Rolls-Royce, we have definitely heard a lot about the emerging sector. Of course this comes as no surprise, as the oft-quoted Gartner prediction goes: there will be over 20 billion connected things by 2020. The increase in connected devices over the past year, from toasters to toothbrushes, shows we may be getting even closer to the forecast. As the opportunity within the IoT sector continues to rise, businesses globally have been taking a leap to developing unique devices or searching for a way to get in on the action with emerging software or platform solutions. Here are some IoT trends to watch out for in 2018, according to the experts.

Summary: Got a good AUC on your hold out data? Think that proves that it's safe to put the model into production. This article shows you some of the pitfalls in this new era of black box Deep Learning Neural Nets and a method for identifying potentially devastating errors. Recently we've been reading that the increased adoption of Deep Learning (DL) image, text, and voice processing tools has been driven by the fact that major developers like Google and Facebook have been able to achieve accuracies of 99% or greater. Deep Learning is indeed on a roll. But before you rush out to add DL to your app, take a step back and consider whether the answers you're getting are in fact right even with very high AUC scores. Deep Learning is an evolution of Neural Nets (NN) and NNs have been around for a long time.

People in Britain are more scared of the artificial intelligence embedded in household devices and self-driving cars than in systems used for predictive policing or diagnosing diseases. That's according to a survey commissioned by the Royal Society, which is billed as the first in-depth look at how the public perceives the risks and benefits associated with machine learning, a key AI technique. Participants in the survey were most worried by the notion that a robot, acting on conclusions derived by machine learning, would cause them physical harm. Accordingly, machines with close proximity to their users, such as those in the home and self-driving cars, were viewed as very risky. The notion of a robot animated by AI is known as "embodiment."