This paper discusses the problem of estimation and inference on time-varying treatments. We propose a method for inference on treatment histories, by introducing a \textit{dynamic} covariate balancing method. Our approach allows for (i) treatments to propagate arbitrarily over time; (ii) non-stationarity and heterogeneity of treatment effects; (iii) high-dimensional covariates, and (iv) unknown propensity score functions. We study the asymptotic properties of the estimator, and we showcase the parametric convergence rate of the proposed procedure. We illustrate in simulations and an empirical application the advantage of the method over state-of-the-art competitors.

Contrastive explanations clarify why an event occurred in contrast to another. They are more inherently intuitive to humans to both produce and comprehend. We propose a methodology to produce contrastive explanations for classification models by modifying the representation to disregard non-contrastive information, and modifying model behavior to only be based on contrastive reasoning. Our method is based on projecting model representation to a latent space that captures only the features that are useful (to the model) to differentiate two potential decisions. We demonstrate the value of contrastive explanations by analyzing two different scenarios, using both high-level abstract concept attribution and low-level input token/span attribution, on two widely used text classification tasks. Specifically, we produce explanations for answering: for which label, and against which alternative label, is some aspect of the input useful? And which aspects of the input are useful for and against particular decisions? Overall, our findings shed light on the ability of label-contrastive explanations to provide a more accurate and finer-grained interpretability of a model's decision.

Dr. LeCun tweets that "ML systems are biased when data is biased." This can be interpreted in multiple ways, with one interpretation being that data is the only factor that matters, and another being that data is the main problem in this particular case. Dr. Gebru replies in an exasperated way noting that the first possible interpretation is incorrect and that experts such as her say this often. Implicitly, it's clear that this exasperation must be partially because this is a common and harmful misconception experts such as Dr. Gebru have to fight against.

In this contributed article, Nic Ray, CEO of BrandsEye, observes that in a post-Covid world, organizations will need to ensure they are using tools that allow them to go beyond simple keyword matching to find the customer conversation that matters. Digital service should by no means spell the demise of the human touch, on the contrary, it reminds us of its importance.

Africa has a unique opportunity to develop its competitiveness through artificial intelligence (AI). From agriculture and remote health to translating the 2,000-odd languages spoken across the continent, AI can help tackle the economic problems that Africa faces. Africa faces several known challenges in developing AI such as a dearth of investment, a paucity of specialised talent, and a lack of access to the latest global research. These hurdles are being whittled down, albeit slowly, thanks to African ingenuity and to investments by multinational companies such as IBM Research, Google, Microsoft, and Amazon, which have all opened AI labs in Africa. Innovative forms of trans-continental collaboration such as Deep Learning Indaba (a Zulu word for gathering), which is fostering a community of AI researchers in Africa, and Zindi, a platform that challenges African data scientists to solve the continent's toughest challenges, are gaining ground, buoyed by the recent "homecoming" of several globally-trained African experts in AI.

The Global Artificial Intelligence for Healthcare Applications market report enumerates highly classified information portfolios encompassing multi-faceted industrial developments with vivid references of market share, size, revenue predictions along with overall regional outlook. The report illustrates a highly dependable overview of the competition isle, with detailed assessment of business verticals. Post a systematic research initiative and subsequent evaluation overview, the global Artificial Intelligence for Healthcare Applications market mimicking its past growth performance is anticipated to strike a flourishing ROI and is therefore more likely to be on the favorable growth curve in the coming years. This versatile report describing the global Artificial Intelligence for Healthcare Applications market has entailed a range of information portfolios that have been segregated into indispensable and additional information streams that have been represented in the form of tables, pie-charts, graphs and the like to align with maximum reader understanding.

The healthcare industry is straining under the impact of COVID-19. The sudden influx of patients in hospitals is exposing vulnerabilities in the current healthcare system. Some hospitals became hotspots for infection, disrupting routine healthcare procedures, while others closed their Outpatient Departments (OPDs), fearing transmission. This dire situation ushered in a massive digital transformation of the healthcare industry to improve care quality, reduce operational costs, and save time for treatments. Although the pandemic accelerated the transformation and saw pioneering research in medical science, healthcare advancement is a phased evolution.

Amidst all the fast-paced technological innovations, contact centres continue to be at the frontline of delivering customer experience. "Even though businesses have identified different mechanisms to reach out to users such as mobile applications, notifications etc, users still reach out to the call center. Case in point, even when you are able to book a cab in under two minutes through the app, you will want to reach out to customer care if there is a problem," shares Sourabh Gupta, Co-Founder & CEO, Vernacular.ai, an AI-first SaaS business enhancing customer experience through intelligent voice conversations. However, Sourabh points out that innovation for contact centres has been overlooked and that's why today they are unable to offer the same convenience that the business provides digitally through other mediums. This gap has come to the fore amidst the pandemic.

To address this challenge, two recent works approach the problem from different perspectives. Tramer et al. (2020) Reliable evaluation of adversarial defenses is a outlines an approach for manually crafting adaptive attacks challenging task, currently limited to an expert that exploit the weak points of each defense. Here, a domain who manually crafts attacks that exploit the defense's expert starts with an existing attack, such as PGD (Madry inner workings, or to approaches based et al., 2018) (denoted as - in Figure 1), and adapts it based on on ensemble of fixed attacks, none of which may knowledge of the defense's inner workings. Common modifications be effective for the specific defense at hand. Our include: (i) tuning attack parameters (e.g., number key observation is that custom attacks are composed of steps), (ii) replacing network components to simplify the from a set of reusable building blocks, attack (e.g., removing randomization or non-differentiable such as fine-tuning relevant attack parameters, network components), and (iii) replacing the loss function optimized transformations, and custom loss functions.

We present a modelling framework for the investigation of supervised learning in non-stationary environments. Specifically, we model two example types of learning systems: prototype-based Learning Vector Quantization (LVQ) for classification and shallow, layered neural networks for regression tasks. We investigate so-called student teacher scenarios in which the systems are trained from a stream of high-dimensional, labeled data. Properties of the target task are considered to be non-stationary due to drift processes while the training is performed. Different types of concept drift are studied, which affect the density of example inputs only, the target rule itself, or both. By applying methods from statistical physics, we develop a modelling framework for the mathematical analysis of the training dynamics in non-stationary environments. Our results show that standard LVQ algorithms are already suitable for the training in non-stationary environments to a certain extent. However, the application of weight decay as an explicit mechanism of forgetting does not improve the performance under the considered drift processes. Furthermore, we investigate gradient-based training of layered neural networks with sigmoidal activation functions and compare with the use of rectified linear units (ReLU). Our findings show that the sensitivity to concept drift and the effectiveness of weight decay differs significantly between the two types of activation function.