Jon a ˇ s Kulh anek 1, Erik Derner 2, Tim de Bruin 1, and Robert Babu ˇ ska 3 Abstract -- Deep reinforcement learning (RL) has been successfully applied to a variety of game-like environments. However, the application of deep RL to visual navigation with realistic environments is a challenging task. We propose a novel learning architecture capable of navigating an agent, e.g. a mobile robot, to a target given by an image. T o achieve this, we have extended the batched A2C algorithm with auxiliary tasks designed to improve visual navigation performance. We propose three additional auxiliary tasks: predicting the segmentation of the observation image and of the target image and predicting the depth-map. These tasks enable the use of supervised learning to pre-train a large part of the network and to reduce the number of training steps substantially. The training performance has been further improved by increasing the environment complexity gradually over time. An efficient neural network structure is proposed, which is capable of learning for multiple targets in multiple environments. Our method navigates in continuous state spaces and on the AI2-THOR environment simulator outperforms state-of-the-art goal-oriented visual navigation methods from the literature. I NTRODUCTION Visual navigation is the problem of navigating an agent, e.g. a mobile robot, in an environment using camera input only. The agent is given a target image (an image it will see from the target position), and its goal is to move from its current position to the target by applying a sequence of actions, based on the camera observations only. We focus on the case when the environment is initially unknown, i.e., no explicit map is available.

Tractable yet expressive density estimators are a key building block of probabilistic machine learning. While sum-product networks (SPNs) offer attractive inference capabilities, obtaining structures large enough to fit complex, high-dimensional data has proven challenging. In this paper, we present random sum-product forests (RSPFs), an ensemble approach for mixing multiple randomly generated SPNs. We also introduce residual links, which reference specialized substructures of other component SPNs in order to leverage the context-specific knowledge encoded within them. Our empirical evidence demonstrates that RSPFs provide better performance than their individual components. Adding residual links improves the models further, allowing the resulting ResSPNs to be competitive with commonly used structure learning methods.

We propose a novel method for explaining the predictions of any classifier. In our approach, local explanations are expected to explain both the outcome of a prediction and how that prediction would change if 'things had been different'. Furthermore, we argue that satisfactory explanations cannot be dissociated from a notion and measure of fidelity, as advocated in the early days of neural networks' knowledge extraction. We introduce a definition of fidelity to the underlying classifier for local explanation models which is based on distances to a target decision boundary. A system called CLEAR: Counterfactual Local Explanations via Regression, is introduced and evaluated. CLEAR generates w-counterfactual explanations that state minimum changes necessary to flip a prediction's classification. CLEAR then builds local regression models, using the w-counterfactuals to measure and improve the fidelity of its regressions. By contrast, the popular LIME method, which also uses regression to generate local explanations, neither measures its own fidelity nor generates counterfactuals. CLEAR's regressions are found to have significantly higher fidelity than LIME's, averaging over 45% higher in this paper's four case studies.

We present a new flavor of Variational Autoencoder (VAE) that interpolates seamlessly between unsupervised, semi-supervised and fully supervised learning domains. We show that unlabeled datapoints not only boost unsupervised tasks, but also the classification performance. Vice versa, every label not only improves classification, but also unsupervised tasks. The proposed architecture is simple: A classification layer is connected to the topmost encoder layer, and then combined with the resampled latent layer for the decoder. The usual evidence lower bound (ELBO) loss is supplemented with a supervised loss target on this classification layer that is only applied for labeled datapoints. This simplicity allows for extending any existing VAE model to our proposed semi-supervised framework with minimal effort. In the context of classification, we found that this approach even outperforms a direct supervised setup.

A trojan backdoor is a hidden pattern typically implanted in a deep neural network. It could be activated and thus forces that infected model behaving abnormally only when an input data sample with a particular trigger present is fed to that model. As such, given a deep neural network model and clean input samples, it is very challenging to inspect and determine the existence of a trojan backdoor. Recently, researchers design and develop several pioneering solutions to address this acute problem. They demonstrate the proposed techniques have a great potential in trojan detection. However, we show that none of these existing techniques completely address the problem. On the one hand, they mostly work under an unrealistic assumption (e.g. assuming availability of the contaminated training database). On the other hand, the proposed techniques cannot accurately detect the existence of trojan backdoors, nor restore high-fidelity trojan backdoor images, especially when the triggers pertaining to the trojan vary in size, shape and position. In this work, we propose TABOR, a new trojan detection technique. Conceptually, it formalizes a trojan detection task as a non-convex optimization problem, and the detection of a trojan backdoor as the task of resolving the optimization through an objective function. Different from the existing technique also modeling trojan detection as an optimization problem, TABOR designs a new objective function--under the guidance of explainable AI techniques as well as heuristics--that could guide optimization to identify a trojan backdoor in a more effective fashion. In addition, TABOR defines a new metric to measure the quality of a trojan backdoor identified. Using an anomaly detection method, we show the new metric could better facilitate TABOR to identify intentionally injected triggers in an infected model and filter out false alarms......

The Federal Trade Commission is hosting a public workshop Wednesday on video game "loot boxes" amid backlash from players claiming it can create a "pay to win" structure. The workshop will bring together industry experts, consumer advocates and others "to discuss concerns regarding the marketing and use of loot boxes and other in-game purchases, and the potential behavioral impact of these virtual rewards on young consumers," according to an event description on the FTC's website. The concept has generated angst among video game players concerned it encourages a "pay to play" atmosphere where gamers must spend money to gain a competitive edge. Loot boxes are rewards players receive within video games containing random prizes. For example, in the Activision Blizzard game Overwatch, loot boxes generate stickers, quotes from characters, and special costumes.

It takes powerful technology to connect our brands and partners with an audience of 1 billion. Nearly half of Verizon Media employees are building the code and platforms that help us achieve that. Whether you're looking to write mobile app code, engineer the servers behind our massive ad tech stacks, or develop algorithms to help us process 4 trillion data points a day, what you do here will have a huge impact on our business--and the world. As Verizon's media unit, our brands like Yahoo, TechCrunch and HuffPost help people stay informed and entertained, communicate and transact, while creating new ways for advertisers and partners to connect. Millions of people visit the Yahoo homepage for news, sports, finance, email, and more.

It takes powerful technology to connect our brands and partners with an audience of 1 billion. Nearly half of Verizon Media employees are building the code and platforms that help us achieve that. Whether you're looking to write mobile app code, engineer the servers behind our massive ad tech stacks, or develop algorithms to help us process 4 trillion data points a day, what you do here will have a huge impact on our business--and the world. As Verizon's media unit, our brands like Yahoo, TechCrunch and HuffPost help people stay informed and entertained, communicate and transact, while creating new ways for advertisers and partners to connect. About Verizon Media Verizon Media is a values-led company committed to building brands people love.

A new analysis shows that the number of Chinese AI researchers has increased tenfold over the last decade, but the majority of them live outside the country. Superpower dreams: China has put forth a concerted effort to grow into a leading AI powerhouse over the last few years. Beijing deemed the discipline in need of special attention as early as 2012, and in 2017 it released a detailed national strategy for advancing and harnessing the technology. Home-grown army: In a new analysis, Joy Dantong Ma, the associate director of MacroPolo, a Chicago-based think tank focused on China's economic growth, showed how this top-down push has affected AI talent. The report analyzed the authorship of papers accepted to NeurIPS, one of the most prestigious international AI conferences, and found a nearly tenfold increase in the number of authors who did their undergraduate studies in China over the last decade.

In fall 2018, the Computing Community Consortium (CCC) initiated an effort to create a 20-Year Roadmap for Artificial Intelligence, led by Yolanda Gil (University of Southern California and President of AAAI) and Bart Selman (Cornell University and President-Elect of AAAI). The goal of the initiative was to identify challenges, opportunities, and pitfalls in the AI landscape, and to create a compelling report to inform future decisions, policies, and investments in this area. The Roadmap was based on broad community input gathered via a number of forums and communication channels: three topical workshops during the fall and winter of 2018/2019, a Town Hall at the annual meeting of the AAAI, and feedback from other groups of stakeholders in industry, government, academia, and the agencies. A draft of the Roadmap was made available on May 13, 2019 for comment by the computing research community. Following the comment period and further revision, the final version was released on August 6, 2019.