Editor's Note: See Joris and Matteo at their tutorial "Opening The Black Box -- Interpretability in Deep Learning" at ODSC Europe 2019 this November 20th in London. In the last decade, the application of deep neural networks to long-standing problems has brought a breakthrough in performance and prediction power. However, high accuracy, deriving from the increased model complexity, often comes at the price of loss of interpretability, i.e., many of these models behave as black-boxes and fail to provide explanations on their predictions. While in certain application fields this issue may play a secondary role, in high-risk domains, e.g., health care, it is crucial to build trust in a model and being able to understand its behavior. The definition of the verb interpret is "to explain or tell the meaning of: present in understandable terms" (Merriam- Webster 2019).

Are you confused by buzzwords like Artificial Intelligence (AI) and DeepLearning? Have you ever wondered if DeepLearning is better than Machine Learning (ML) or why AI and ML are used interchangeably? Keep reading to decode these tech-jargons. As technologies rise in the Hype-cycle towards their "Peak of Inflated Expectations", they tend to loose their precise technical definitions. This is because non-partitioners introduce a more colloquial interpretation, which can confuse the space.

Convolutional Neural Networks(CNNs) and other deep learning networks have enabled unprecedented breakthroughs in a variety of computer vision tasks from image classification to object detection, semantic segmentation, image captioning and more recently visual question answering. While these networks enable superior performance, their lack of decomposability into intuitive and understandable components makes them hard to interpret. Consequently, when today's intelligent systems fail, they fail spectacularly disgracefully without warning or explanation, leaving a user staring at an incoherent output, wondering why. Interpretability of Deep Learning models matters to build trust and move towards their successful integration in our daily lives. To achieve this goal the model transparency is useful to explain why they predict what they predict.

PyTorch has debuted a slew of experimental features in its just-released version 1.3 as support for the TensorFlow competitor broadens, and new tools to tackle challenges like privacy appear. PyTorch 1.3 seems to be right on trend with its new capabilities, adding, for example, previews of implementations for model quantisation and on-device machine learning. The latter is heavily looked into these days, as interest in privacy-focused approaches soars. Mobile support is one of the building blocks to, for example, realise federated learning, a technique which allows training data to be spread between clients, meaning that data doesn't have to leave a device anymore to be included in the training of a centralised model. In its first iteration, mobile support comes down to prebuilt LibTorch libraries for Android and iOS, optimised implementations for certain operators, modules making sure that TorchScript inference is possible and forward operations can be executed on mobile CPUs.

It's often assumed that as the complexity of an AI system increases, it becomes invariably less interpretable. But researchers have begun to challenge that notion with libraries like Facebook's Captum, which explains decisions made by neural networks with the deep learning framework PyTorch, as well as IBM's AI Explainability 360 toolkit and Microsoft's InterpretML. In a bid to render AI's decision-making even more transparent, a team hailing from Google and Stanford recently explored a machine learning model -- Automated Concept-based Explanation (ACE) -- that automatically extracts the "human-meaningful" visual concepts informing a model's predictions. As the researchers explain in a paper detailing their work, most machine learning explanation methods alter individual features (e.g., pixels, super-pixels, word-vectors) to approximate the importance of each to the target model. This is an imperfect approach, because it's vulnerable to even the smallest shifts in the input.

It is no secret that deep neural networks (DNNs) can achieve state-of-the-art performance in a wide range of complicated tasks. DNN models such as BigGAN, BERT, and GPT 2.0 have proved the high potential of deep learning. Deploying DNNs on mobile devices, consumer devices, drones and vehicles however remains a bottleneck for researchers. For such practical, on-device scenarios, DNNs must have a smaller footprint. The requirement for smaller DNNs has pushed researchers in two opposite directions: either hand-craft DNNs through design principles, or rely entirely on automated network architecture search.

Conventional, data-crunching artificial intelligence, which is the foundation of deep learning, isn't enough on its own; the human-like reasoning of symbolic artificial intelligence is fascinating, but on its own, it isn't enough either. The unique hybrid combination of the two -- numeric data analytics techniques that include statistical analysis, modeling, and machine learning, plus the explainability (and transparency) of symbolic artificial intelligence -- is now termed "cognitive AI." It's an extraordinary breakthrough to have the ability to implement a human-like ability to perceive, understand, correlate, learn, teach, reason, and solve problems faster than existing AI solutions. Key technology components were at the core of the wildly successful NASA Mars Rover's mission. Alone and 150 million miles from Earth, the rover was able to successfully adapt to conditions without direct instruction. After a dust storm, it taught itself to rotate its solar panels and shake off accumulated dust blocking essential solar ray absorption.

Effective and efficient recommendation is crucial for modern e-commerce platforms. It consists of two indispensable components named Click-Through Rate (CTR) prediction and Conversion Rate (CVR) prediction, where the latter is an essential factor contributing to the final purchasing volume. Existing methods specifically predict CVR using the clicked and purchased samples, which has limited performance affected by the well-known sample selection bias and data sparsity issues. To address these issues, we propose a novel deep CVR prediction method by considering the post-click behaviors. After grouping deterministic actions together, we construct a novel sequential path, which elaborately depicts the post-click behaviors of users. Based on the path, we define the CVR and several related probabilities including CTR, etc., and devise a deep neural network with multiple targets involved accordingly. It takes advantage of the abundant samples with deterministic labels derived from the post-click actions, leading to a significant improvement of CVR prediction. Extensive experiments on both offline and online settings demonstrate its superiority over representative state-of-the-art methods.

Improving Robustness of time series classifier with Neural ODE guided gradient based data augmentation Anindya Sarkar Mobiliya Bangalore, INDIA anindya.sarkar@mobiliya.com Abstract --Exploring adversarial attack vectors and studying their effects on machine learning algorithms has been of interest to researchers. Deep neural networks working with time series data have received lesser interest compared to their image counterparts in this context. In a recent finding, it has been revealed that current state-of-the-art deep learning time series classifiers are vulnerable to adversarial attacks. In this paper, we introduce two local gradient based and one spectral density based time series data augmentation techniques. We show that a model trained with data obtained using our techniques obtains state-of- the-art classification accuracy on various time series benchmarks. In addition, it improves the robustness of the model against some of the most common corruption techniques,such as Fast Gradient Sign Method (FGSM) and Basic Iterative Method (BIM). Index T erms --time series classification, adversarial training, gradient based adversarial attacks I.

Flooding is a destructive and dangerous hazard and climate change appears to be increasing the frequency of catastrophic flooding events around the world. Physics-based flood models are costly to calibrate and are rarely generalizable across different river basins, as model outputs are sensitive to site-specific parameters and human-regulated infrastructure. In contrast, statistical models implicitly account for such factors through the data on which they are trained. Such models trained primarily from remotely-sensed Earth observation data could reduce the need for extensive in-situ measurements. In this work, we develop generalizable, multi-basin models of river flooding susceptibility using geographically-distributed data from the USGS stream gauge network. Machine learning models are trained in a supervised framework to predict two measures of flood susceptibility from a mix of river basin attributes, impervious surface cover information derived from satellite imagery, and historical records of rainfall and stream height. We report prediction performance of multiple models using precision-recall curves, and compare with performance of naive baselines. This work on multi-basin flood prediction represents a step in the direction of making flood prediction accessible to all at-risk communities.