Google's latest Doodle celebrates the life of Michael Dertouzos, the Greek computer scientist who anticipated how the internet would come to dominate almost every aspect of our lives. Monday's homepage image honoured the renowned academic, who died in 2001, on what would have been his 82nd birthday. As the director of MIT's Laboratory for Computer Science for almost 30 years, Mr Dertouzos was a pivotal figure in the creation of the World Wide Web Consortium – the alliance of companies and groups promoting the evolution of the internet. Canadian Google Doodle celebrates dino-hunting Joseph Burr Tyrrell How to play the ghostly new Google Doodle game Google Doodle reveals the most searched for recipes in every state Google Doodle celebrates female artists on International Women's Day Google Doodle celebrates female artists on International Women's Day Mr Dertouzos even recruited Tim Berners-Lee, the primary inventor of the World Wide Web, to lead the consortium. Mr Berners-Lee said: "If it hadn't been for Michael there would not probably have been a World Wide Web Consortium." Born in Athens in 1936, Mr Dertouzos won a Fulbright scholarship to the University of Arkansas.

Superrealistic plastic face masks produced by a firm in Otsu, Shiga Prefecture, have recently attracted attention at home and abroad, from facial-recognition system developers to a Saudi Arabian royal family member. "Look, it makes your heart pound, doesn't it?" The masks -- named Real Face -- are made of plastic resin roughly 1 to 2 millimeters thick. Kitagawa says he came up with the idea of making realistic masks more than a decade ago, when he was developing copy machines at a major printing device manufacturer. "I wanted to make copies of human beings," he said.

Those who still think China has a never-ending army of cheap labor should think again. Chinese wages are rising, the population is aging - and China's government is promoting a robot revolution in manufacturing. Over each of the next three years, the growth of the use of robots in China is estimated to exceed 20 percent. Rising wages - driven by decades of growth - are eating into profits, pushing companies to shift manufacturing to Southeast Asia. Shanghai's minimum monthly wage is two-and-a-half times what it was a decade ago.

History has time and again taught us that, science fiction is only a fantasy until science makes it a reality. In the 1940s Isaac Asimov, a prolific science fiction writer wrote about a future where robots are a part of the human world. Similarly, in a sci-fi film, Robocop made more than 30 years ago, a robot is built-up in order to solve an unprecedented crime problem in dystopian crime-ridden Detroit. Today science fiction has become a reality. Police in different parts of the world are using robots for law enforcement and first, ever robotic police officers have become deployed across China, Dubai and Hyderabad in India.

Imagine a country with an army of techies, a government that supports AI and blockchain by setting a mandate and investing billions, large scale tech companies that are rapidly experimenting and implementing at scale, and an abundance of data to feed the application of these technologies. This just about covers the AI and blockchain playground that is China. The gloves are off, and over the coming years some of the greatest advancements will emanate from the east. An ambitious AI strategic plan was laid by the China's State Council in July 2017, aiming to create a domestic 1 trillion yuan ($150 billion) AI industry by 2030. Following this, Chinese president Xi Jinping called upon his country to take the lead in developing new technologies like artificial intelligence, the internet of things, and blockchain.

Towards a Near Universal Time Series Data Mining Tool: Introducing the Matrix Profile by Chin-Chia Michael Yeh Doctor of Philosophy, Graduate Program in Computer Science University of California, Riverside, September 2018 Dr. Eamonn Keogh, Chairperson The last decade has seen a flurry of research on all-pairs-similarity-search (or, self-join) for text, DNA, and a handful of other datatypes, and these systems have been applied to many diverse data mining problems. Surprisingly, however, little progress has been made on addressing this problem for time series subsequences. In this thesis, we have introduced a near universal time series data mining tool called matrix profile which solves the all-pairssimilarity-search problem and caches the output in an easy-to-access fashion. The proposed algorithm is not only parameter-free, exact and scalable, but also applicable for both single and multidimensional time series. By building time series data mining methods on top of matrix profile, many time series data mining tasks (e.g., motif discovery, discord discovery, shapelet discovery, semantic segmentation, and clustering) can be efficiently solved. Because the same matrix profile can be shared by a diverse set of time series data mining methods, matrix profile is versatile and computed-once-use-many-times data structure. We demonstrate the utility of matrix profile for many time series data mining problems, including motif discovery, discord discovery, weakly labeled time series classification, and vi representation learning on domains as diverse as seismology, entomology, music processing, bioinformatics, human activity monitoring, electrical power-demand monitoring, and medicine. We hope the matrix profile is not the end but the beginning of many more time series data mining projects.

We develop the mathematical foundations of the stochastic modified equations (SME) framework for analyzing the dynamics of stochastic gradient algorithms, where the latter is approximated by a class of stochastic differential equations with small noise parameters. We prove that this approximation can be understood mathematically as an weak approximation, which leads to a number of precise and useful results on the approximations of stochastic gradient descent (SGD), momentum SGD and stochastic Nesterov's accelerated gradient method in the general setting of stochastic objectives. We also demonstrate through explicit calculations that this continuous-time approach can uncover important analytical insights into the stochastic gradient algorithms under consideration that may not be easy to obtain in a purely discrete-time setting. Keywords: stochastic gradient algorithms, modified equations, stochastic differential equations, momentum, Nesterov's accelerated gradient

Most modern neural machine translation (NMT) systems rely on presegmented inputs. Segmentation granularity importantly determines the input and output sequence lengths, hence the modeling depth, and source and target vocabularies, which in turn determine model size, computational costs of softmax normalization, and handling of out-of-vocabulary words. However, the current practice is to use static, heuristic-based segmentations that are fixed before NMT training. This begs the question whether the chosen segmentation is optimal for the translation task. To overcome suboptimal segmentation choices, we present an algorithm for dynamic segmentation based on the Adaptative Computation Time algorithm (Graves 2016), that is trainable end-to-end and driven by the NMT objective. In an evaluation on four translation tasks we found that, given the freedom to navigate between different segmentation levels, the model prefers to operate on (almost) character level, providing support for purely character-level NMT models from a novel angle.

Making learners robust to adversarial perturbation at test time (i.e., evasion attacks) or training time (i.e., poisoning attacks) has emerged as a challenging task. It is known that for some natural settings, sublinear perturbations in the training phase or the testing phase can drastically decrease the quality of the predictions. These negative results, however, are information theoretic and only prove the existence of such successful adversarial perturbations. A natural question for these settings is whether or not we can make classifiers computationally robust to polynomial-time attacks. In this work, we prove strong barriers against achieving such envisioned computational robustness both for evasion and poisoning attacks. In particular, we show that if the test instances come from a product distribution (e.g., uniform over $\{0,1\}^n$ or $[0,1]^n$, or isotropic $n$-variate Gaussian) and that there is an initial constant error, then there exists a polynomial-time attack that finds adversarial examples of Hamming distance $O(\sqrt n)$. For poisoning attacks, we prove that for any learning algorithm with sample complexity $m$ and any efficiently computable "predicate" defining some "bad" property $B$ for the produced hypothesis (e.g., failing on a particular test) that happens with an initial constant probability, there exist polynomial-time online poisoning attacks that tamper with $O (\sqrt m)$ many examples, replace them with other correctly labeled examples, and increases the probability of the bad event $B$ to $\approx 1$. Both of our poisoning and evasion attacks are black-box in how they access their corresponding components of the system (i.e., the hypothesis, the concept, and the learning algorithm) and make no further assumptions about the classifier or the learning algorithm producing the classifier.

As statistical classifiers become integrated into real-world applications, it is important to consider not only their accuracy but also their robustness to changes in the data distribution. Although identifying and controlling for confounding variables Z - correlated with both the input X of a classifier and its output Y - has been assiduously studied in empirical social science, it is often neglected in text classification. This can be understood by the fact that, if we assume that the impact of confounding variables does not change between the time we fit a model and the time we use it, then prediction accuracy should only be slightly affected. We show in this paper that this assumption often does not hold and that when the influence of a confounding variable changes from training time to prediction time (i.e. under confounding shift), the classifier accuracy can degrade rapidly. We use Pearl's back-door adjustment as a predictive framework to develop a model robust to confounding shift under the condition that Z is observed at training time. Our approach does not make any causal conclusions but by experimenting on 6 datasets, we show that our approach is able to outperform baselines 1) in controlled cases where confounding shift is manually injected between fitting time and prediction time 2) in natural experiments where confounding shift appears either abruptly or gradually 3) in cases where there is one or multiple confounders. Finally, we discuss multiple issues we encountered during this research such as the effect of noise in the observation of Z and the importance of only controlling for confounding variables.