AI robots are still not sophisticated enough to understand humans or the complexity of social situations, says UNSW's Dr Masimiliano Cappuccio. "So we need to think about how we interact with social and companion robots to instead help us become more aware of our own behaviour, limitations, vices or bad habits," says Dr Cappuccio, the Deputy Director of Values in Defense and Security Technology at UNSW Canberra. "And this can be in the areas of greater self-discipline and self-control but also in learning virtues such as generosity and empathy." Dr Cappuccio is the lead author of Can Robots Make Us Better Humans? Virtuous Robotics and the Good Life with Artificial Agents which was written in collaboration with UNSW Art & Design's Dr Eduardo Sandoval and Professor Mari Velonaki along with academics from the University of Western Sydney and Chalmers University of Technology in Sweden It is also the first in a collection co-edited by Dr Cappuccio, Dr Sandoval and Prof. Velonaki and published in the International Journal of Robotics as a special issue titled Virtuous Robotics: Artificial Agents and the Good Life.

Australia Post is using machine learning to calculate the arrival time of parcels down to a two-hour window based on data from a new route planner that has just been rolled out to delivery drivers. Executive general manager for transformation and enablement John Cox on Tuesday said the new feature is currently being trialled to give customers a more accurate estimate on delivery times. "What we're trialling – and this is not out in the public yet – is what we call an estimated time of arrival," Cox told the Digital Transformation Agency's 2020 Digital Summit. "So based off when the postie scans the parcel in the morning to put in their van, we'll be able to notify the consumer that it will be delivered within that window of time." The estimate is calculated using the "single scanning platform" that all deliver drivers use to scan their parcels before their run, which also determines the "optimal" route for each delivery run.

We develop a general framework for constructing new information-theoretic divergences that rigorously interpolate between $f$-divergences and integral probability metrics (IPMs), such as the Wasserstein distance. These new divergences inherit features from IPMs, such as the ability to compare distributions which are not absolute continuous, as well as from $f$-divergences, for instance the strict concavity of their variational representations and the ability to compare heavy-tailed distributions. When combined, these features establish a divergence with improved convergence and estimation properties for statistical learning applications. We demonstrate their use in the training of generative adversarial networks (GAN) for heavy-tailed data and also show they can provide improved performance over gradient-penalized Wasserstein GAN in image generation.

In this paper, we study the Empirical Risk Minimization (ERM) problem in the non-interactive Local Differential Privacy (LDP) model. Previous research on this problem \citep{smith2017interaction} indicates that the sample complexity, to achieve error $\alpha$, needs to be exponentially depending on the dimensionality $p$ for general loss functions. In this paper, we make two attempts to resolve this issue by investigating conditions on the loss functions that allow us to remove such a limit. In our first attempt, we show that if the loss function is $(\infty, T)$-smooth, by using the Bernstein polynomial approximation we can avoid the exponential dependency in the term of $\alpha$. We then propose player-efficient algorithms with $1$-bit communication complexity and $O(1)$ computation cost for each player. The error bound of these algorithms is asymptotically the same as the original one. With some additional assumptions, we also give an algorithm which is more efficient for the server. In our second attempt, we show that for any $1$-Lipschitz generalized linear convex loss function, there is an $(\epsilon, \delta)$-LDP algorithm whose sample complexity for achieving error $\alpha$ is only linear in the dimensionality $p$. Our results use a polynomial of inner product approximation technique. Finally, motivated by the idea of using polynomial approximation and based on different types of polynomial approximations, we propose (efficient) non-interactive locally differentially private algorithms for learning the set of k-way marginal queries and the set of smooth queries.

The complete blood count (CBC) performed by automated hematology analyzers is one of the most ordered laboratory tests. It is a first-line tool for assessing a patient's general health status, or diagnosing and monitoring disease progression. When the analysis does not fit an expected setting, technologists manually review a blood smear using a microscope. The International Consensus Group for Hematology Review published in 2005 a set of criteria for reviewing CBCs. Commonly, adjustments are locally needed to account for laboratory resources and populations characteristics. Our objective is to provide a decision support tool to identify which CBC variables are associated with higher risks of abnormal smear and at which cutoff values. We propose a cost-sensitive Lasso-penalized additive logistic regression combined with stability selection. Using simulated and real CBC data, we demonstrate that our tool correctly identify the true cutoff values, provided that there is enough available data in their neighbourhood.

Inspired by the success of deep learning, recent industrial Click-Through Rate (CTR) prediction models have made the transition from traditional shallow approaches to deep approaches. Deep Neural Networks (DNNs) are known for its ability to learn non-linear interactions from raw feature automatically, however, the non-linear feature interaction is learned in an implicit manner. The non-linear interaction may be hard to capture and explicitly model the \textit{co-action} of raw feature is beneficial for CTR prediction. \textit{Co-action} refers to the collective effects of features toward final prediction. In this paper, we argue that current CTR models do not fully explore the potential of feature co-action. We conduct experiments and show that the effect of feature co-action is underestimated seriously. Motivated by our observation, we propose feature Co-Action Network (CAN) to explore the potential of feature co-action. The proposed model can efficiently and effectively capture the feature co-action, which improves the model performance while reduce the storage and computation consumption. Experiment results on public and industrial datasets show that CAN outperforms state-of-the-art CTR models by a large margin. Up to now, CAN has been deployed in the Alibaba display advertisement system, obtaining averaging 12\% improvement on CTR and 8\% on RPM.

In this work, we present an ensemble of descriptors for the classification of transmission electron microscopy images of viruses. We propose to combine handcrafted and deep learning approaches for virus image classification. The set of handcrafted is mainly based on Local Binary Pattern variants, for each descriptor a different Support Vector Machine is trained, then the set of classifiers is combined by sum rule. The deep learning approach is a densenet201 pretrained on ImageNet and then tuned in the virus dataset, the net is used as features extractor for feeding another Support Vector Machine, in particular the last average pooling layer is used as feature extractor. Finally, classifiers trained on handcrafted features and classifier trained on deep learning features are combined by sum rule. The proposed fusion strongly boosts the performance obtained by each stand-alone approach, obtaining state of the art performance.

In tis paper we consider approaches for time series forecasting based on deep neural networks and neuro-fuzzy nets. Also, we make short review of researches in forecasting based on various models of ANFIS models. Deep Learning has proven to be an effective method for making highly accurate predictions from complex data sources. Also, we propose our models of DL and Neuro-Fuzzy Networks for this task. Finally, we show possibility of using these models for data science tasks. This paper presents also an overview of approaches for incorporating rule-based methodology into deep learning neural networks.

A team of scientists built a tentacle-like robot gripper that they say is more sensitive than a conventional claw or hand-shaped machine -- because it wraps around and delicately constricts objects like a python or an elephant's trunk. Modeling an elephant's complex musculature allows it to snugly grip delicate or hard-to-reach objects, according to a press release, in a new, gentler paradigm for robotic arms. Thanks to a network of sensors that tell the robot snake how hard it's gripping, the University of New South Wales (UNSW) scientists that designed it say it can grab and hold delicate objects far more safely than conventional robots, according to research published last week in the journal Advanced Materials Technologies. In a short lab video, the robot can be seen automatically winding around and grabbing objects ranging from a hammer to cucumber. The scientists say it can hold a grip on objects 220 times heavier than itself.

Chris Johnston, Executive Director for Paerata Rise says "Our goal is to be one of the most desirable places to live in Auckland and becoming a Smart Village is an extension of this. It means we are able to offer the utmost connection to our residents through a private network, and the most cutting-edge technologies." "We are proud to be able to support the unique, first class community being built at Paerata Rise with the latest concepts in mobile network technology. While everyone has experienced the frustration of bad coverage, ultimately an excellent network should go unnoticed – instead allowing users' mobile applications, services, and other benefits to come to the fore," adds Ross Spearman, General Manager of Dense Air New Zealand. "The realities of smart villages and connected neighbourhoods are starting to emerge in the world around us, however the real'smarts' need to be built into the foundations of communities which is why we are executing this project at this stage of our development," says Johnston.