It wasn't that long ago that Sophia Withers was travelling around Australia and Asia after her undergraduate degree. After a stint working on farms and in brunch cafes, she moved to Melbourne and joined KPMG Australia. While working as a coordinator for their audit division, she began studying the University of Birmingham's part-time Online MSc International Business, in 2018. It was during the program that she deep dived into her passion for blockchain and emerging technology. Her degree research project--How will Blockchain 3.0 facilitate social and economic impact?

Today's marketing teams are filled with specialists in fields such as data analytics or the use of specific tools, and the introduction of artificial intelligence (AI) is threatening to increase the need for dedicated professionals. For marketing leaders this has created the requirement to develop capabilities in areas that might be far removed from why they got into marketing in the first place. But perhaps AI might also be the tool that finally lets marketers get back to doing marketing? This is the promise of recently born company, Acoustic, which brings together many of the cloud-based marketing products and technologies formerly housed within IBM. According to Acoustic's senior vice-president of product management, Jay Henderson, Acoustic is taking advantage of IBM's heritage as one of the early adopters of AI within marketing.

GNW The global artificial intelligence (AI) in retail market attained $720.0 million in 2018 and is predicted to witness a CAGR of 35.4% during 2019–2024 (forecast period). The factors contributing to the growth of the market include the increasing investments in AI by retail companies and expanding e-retail industry. With AI, retailers have been able to automate their work processes, study consumer behavior, and capture relevant data through the adoption of numerous advanced technologies, such as machine learning, natural language processing (NLP), and computer vision. When technology is considered, the AI in retail market is divided into computer vision, NLP, machine learning, and others (which include gesture recognition and analytics). Machine learning generated the highest revenue during the historical period (2014–2018) and is expected to dominate the market during the forecast period as well.

The Australian construction industry is far behind the rest of the world when it comes to digital innovation – a critical ingredient in meeting the demands of our rapidly growing population. With immense building and infrastructure pipelines in our cities, we must, as a nation and an industry, embrace digital innovation and artificial intelligence in order to unlock greater efficiencies, improved productivity and accuracies to future-proof our cities, particularly as we face scarcity in labour resources. Infrastructure planning and delivery requires a very specific skillset, not one that is easily transferable from the building sector. It should be home grown through our university system and imported from overseas through skilled migrants as our general population and transportation requirements are growing so quickly that we can't keep up. Simply, there's more work than there are skilled workers.

Currently, the digital media is in a transitional phase, where the format of the medium is changing from text-based to one with visuals. Due to this significant shift, advertising has to play catch up, to stay up-to-date with the latest trends the industry. On top of that, the marketing industry has to deal with ad-blockers, which blocks out intrusive advertisements. According to a study done by PageFair, there are at least 615 million devices that use Adblock regularly. As you can imagine, getting through these ad-blockers is an uphill task, because they keep disruptive advertisements at bay.

Noname manuscript No. (will be inserted by the editor)Time series classification for varying length series Chang Wei Tan · Fran cois Petitjean· Eamonn Keogh · Geoffrey I. Webb the date of receipt and acceptance should be inserted later Abstract Research into time series classification has tended to focus on the case of series of uniform length. However, it is common for real-world time series data to have unequal lengths. Differing time series lengths may arise from a number of fundamentally different mechanisms. In this work, we identify and evaluate two classes of such mechanisms - variations in sampling rate relative to the relevant signal and variations between the start and end points of one time series relative to one another. We investigate how time series generated by each of these classes of mechanism are best addressed for time series classification. We perform extensive experiments and provide practical recommendations on how variations in length should be handled in time series classification. Keywords Time Series Classification, Proximity Forest, Dynamic Time Warping 1 Introduction Time series classification (TSC) is an important task in many modern world applications such as remote sensing (Pelletier et al., 2019; Petitjean et al., 2012), astronomy (Batista et al., 2011), speech recognition (Hamooni et al., 2016), and insect classification (Chen et al., 2014). The time series to be classified are the observed outputs generated by some process. The classification task often relates to identifying the class of the process that generated the series. Each class of process might be considered as a realization of one or more ideals (in the Platonic sense) or prototypes. The resulting time series can then beChang Wei Tan · Fran cois Petitjean· Geoffrey I. Webb Faculty of Information Technology 25 Exhibition Walk Monash University, Melbourne VIC 3800, Australia Email: chang.tan@monash.edu,francois.petitjean@monash.edu,geoff.webb@monash.edu An observed time series might differ from the ideal in many ways. Much of the research on time series distance measures in the last decade can be seen as the introduction of techniques to mitigate these differences, either as a preprocessing step or directly in a distance measure. For example, variations in amplitude and offset are typically addressed in time series classification by normalization of the series (Rakthanmanon et al., 2012). Some observed values may be erroneous and might be addressed by outlier detection (Basu and Meckesheimer, 2007) and subsequent reinterpolation (Pelletier et al., 2019).

Instead, we propose to integrate a generative inpainter into three representative attribution methods to remove an input feature. Compared to the original counterparts, our methods (1) generate more plausible counterfactual samples under the true data generating process; (2) are more robust to hyperparameter settings; and (3) localize objects more accurately. Our findings were consistent across both ImageNet and Places365 datasets and two different pairs of classifiers and inpainters. Explaining a classifier's outputs given a certain input is increasingly important, especially for life-critical applications (Doshi-V elez & Kim, 2017). A popular means for visually explaining an image classifier's decisions is an attribution map i.e. a heatmap that highlights the input pixels that are the evidence for and against the classification outputs (Montavon et al., 2018). To construct an attribution map, many methods approximate the attribution value of an input region by the classification probability change when that region is absent i.e. removed from the image. That is, most perturbation-based attribution methods implement the absence of an input feature by replacing it with (a) mean pixels; (b) random noise; or (c) blurred versions of the original content. While removing an input feature to measure its attribution is a principle method in causal reasoning, the existing removal (i.e. To combat these two issues, we propose to harness a state-of-the-art generative inpainting model (hereafter, an inpainter) to remove features from an input image and fill in with content that is plausible under the true data distribution. We test our approach on three representative attribution methods of Sliding-Patch (SP) (Zeiler & Fergus, 2014), LIME (Ribeiro et al., 2016), and Meaningful-Perturbation (MP) (Fong & V edaldi, 2017) across two large-scale datasets of ImageNet (Russakovsky et al., 2015) and Places365 (Zhou et al., 2017). For each dataset, we use a separate pair of pre-trained image classifiers and inpainters. Work done during CA's internship at Auburn University.

Adaptive optimization algorithms such as Adam (Kingma & Ba, 2014) are widely used in deep learning. The stability of such algorithms is often improved with a warmup schedule for the learning rate. Motivated by the difficulty of choosing and tuning warmup schedules, Liu et al. (2019) propose automatic variance rectification of Adam's adaptive learning rate, claiming that this rectified approach ("RAdam") surpasses the vanilla Adam algorithm and reduces the need for expensive tuning of Adam with warmup. In this work, we point out various shortcomings of this analysis. We then provide an alternative explanation for the necessity of warmup based on the magnitude of the update term, which is of greater relevance to training stability. Finally, we provide some "rule-of-thumb" warmup schedules, and we demonstrate that simple untuned warmup of Adam performs more-or-less identically to RAdam in typical practical settings. We conclude by suggesting that practitioners stick to linear warmup with Adam, with a sensible default being linear warmup over $2 / (1 - \beta_2)$ training iterations.

Recent policy optimization approaches have achieved substantial empirical success by constructing surrogate optimization objectives. The Approximate Policy Iteration objective (Schulman et al., 2015a; Kakade and Langford, 2002) has become a standard optimization target for reinforcement learning problems. Using this objective in practice requires an estimator of the advantage function. Policy optimization methods such as those proposed in Schulman et al. (2015b) estimate the advantages using a parametric critic. In this work we establish conditions under which the parametric approximation of the critic does not introduce bias to the updates of surrogate objective. These results hold for a general class of parametric policies, including deep neural networks. We obtain a result analogous to the compatible features derived for the original Policy Gradient Theorem (Sutton et al., 1999). As a result, we also identify a previously unknown bias that current state-of-the-art policy optimization algorithms (Schulman et al., 2015a, 2017) have introduced by not employing these compatible features.

Recent policy optimization approaches (Schulman et al., 2015a, 2017) have achieved substantial empirical successes by constructing new proxy optimization objectives. These proxy objectives allow stable and low variance policy learning, but require small policy updates to ensure that the proxy objective remains an accurate approximation of the target policy value. In this paper we derive an alternative objective that obtains the value of the target policy by applying importance sampling. This objective can be directly estimated from samples, as it takes an expectation over trajectories generated by the current policy. However, the basic importance sampled objective is not suitable for policy optimization, as it incurs unacceptable variance. We therefore introduce an approximation that allows us to directly trade-off the bias of approximation with the variance in policy updates. We show that our approximation unifies the proxy optimization approaches with the importance sampling objective and allows us to interpolate between them. We then provide a theoretical analysis of the method that directly quantifies the error term due to the approximation. Finally, we obtain a practical algorithm by optimizing the introduced objective with proximal policy optimization techniques (Schulman etal., 2017). We empirically demonstrate that the result-ing algorithm yields superior performance on continuous control benchmarks