Machine learning and AI systems need data to function, but they also need to be actively protected. IBM researcher Dr Irina Nicolae is applying her skills to these complex issues. Dr Irina Nicolae is a research scientist at IBM Research Ireland. With a background in computer science and software engineering and a PhD in machine learning (ML), Nicolae has carved out a career battling one of security's most pressing issues: protecting artificial intelligence (AI) and ML systems from attacks. The quality of the data and its relevance to the task are the most important points to consider.

What is the future of mobility? According to NAVYA, an innovative French company, it is both autonomous and electric. The benefits of such technology in terms of mobility are incredible. Whilst using electric propulsion reduces CO2 emissions, autonomous driving should reduce traffic congestion and could also be an innovative way to have more efficient public transportation. It is quite a busy moment for NAVYA, who became listed on the stock market last summer.

The banking industry is becoming a digital rather than a physical system. So what sort of leaders should be running a modern bank? Should they be accountants or engineers -- or both? When I was completing a PhD in artificial intelligence (AI) at the University of Cambridge 20 years ago, many of my engineering classmates went to work for banks, and some of them run large Wall Street financial companies. The fourth industrial revolution is shrinking the world of work at a rapid rate.

Paul Duan was working as a data scientist at Eventbrite in San Francisco by day, and volunteering at homeless shelters and soup kitchens by night. He realized one day that he wanted to use AI to help unemployed people find jobs--a core mission of his Paris, France-based nonprofit Bayes Impact. Bayes Impact uses data to build social services fit for a better future. "When you work at a soup kitchen, you serve a soup one by one to each individual, and it feels great," says Duan, "But then it gets really sad because you see that there are 50 people in line behind the person, and you know that behind the closed door of the shelter you have 10,000 more on the streets. So the one question that came to mind was, 'how can we impact people at the biggest scale?'"

It is designed to plug in to Metacorps, an experimental web app for annotating metaphor. As Metacorps users annotate metaphors, the system will use user annotations as training data. When the system is confident, it will suggest an identification and an annotation. Once approved by the user, this becomes more training data. This naturally allows for transfer learning, where the system can, with some known degree of reliability, classify one class of metaphor after only being trained on another class of metaphor. For example, in our metaphorical violence project, metaphors may be classified by the network they were observed on, the grammatical subject or object of the violence metaphor, or the violent word used (hit, attack, beat, etc.).

In this paper, we consider the following problem: a group of robots have to explore a labyrinthine environment and find an intruder if need be. Such a problem has been already studied in mainly two different aspects. On the one hand, the "prey predators" problem has been proposed for the first time by [BJD86]: the prey and the predators, i.e. the agents, share a common environment which is represented by a mere grid. The hunt is simulated and the agents that are autonomous processes can move horizontally or vertically all the time. Quite obviously, the goal of the prey is to escape as long as possible whereas the predators have to capture the prey as fast as possible. Therefore, the "prey predators" problem is an interesting testbed for competing agents and coordination protocols. Indeed, because each agent does not perceive all the environment, and can have incomplete or inconsistent knowledge about the other agents, many challenging issues have to be tackled: which information has to be communicated?

Recently, due to the increasing popularity of social media, the necessity for extracting information from informal text types, such as microblog texts, has gained significant attention. In this study, we focused on the Named Entity Recognition (NER) problem on informal text types for Turkish. We utilized a semi-supervised learning approach based on neural networks. We applied a fast unsupervised method for learning continuous representations of words in vector space. We made use of these obtained word embeddings, together with language independent features that are engineered to work better on informal text types, for generating a Turkish NER system on microblog texts. We evaluated our Turkish NER system on Twitter messages and achieved better F-score performances than the published results of previously proposed NER systems on Turkish tweets. Since we did not employ any language dependent features, we believe that our method can be easily adapted to microblog texts in other morphologically rich languages.

This paper introduces PyDCI, a new implementation of Distributional Correspondence Indexing (DCI) written in Python. DCI is a transfer learning method for cross-domain and cross-lingual text classification for which we had provided an implementation (here called JaDCI) built on top of JaTeCS, a Java framework for text classification. PyDCI is a stand-alone version of DCI that exploits scikit-learn and the SciPy stack. We here report on new experiments that we have carried out in order to test PyDCI, and in which we use as baselines new high-performing methods that have appeared after DCI was originally proposed. These experiments show that, thanks to a few subtle ways in which we have improved DCI, PyDCI outperforms both JaDCI and the above-mentioned high-performing methods, and delivers the best known results on the two popular benchmarks on which we had tested DCI, i.e., MultiDomainSentiment (a.k.a. MDS -- for cross-domain adaptation) and Webis-CLS-10 (for cross-lingual adaptation). PyDCI, together with the code allowing to replicate our experiments, is available at https://github.com/AlexMoreo/pydci .

In many applications of statistics and machine learning, we wish to learn models that achieve uniformly good performance over almost all input values. This is important for safety-and fairnesscritical systems such as medical diagnosis, autonomous vehicles, criminal justice and credit evaluations, where poor performance on the tails of the inputs leads to high-cost system failures. Methods that optimize average performance, however, often produce models that suffer low performance on the "hard" instances of the population. For example, standard regressors obtained from maximum likelihood estimation can lose their predictive power on certain regions of covariates [57], so that high average performance comes at the expense of low performance on minority subpopulations. In this work, we propose and study a procedure that explicitly optimizes performance on tail inputs that suffer high loss. Modern datasets incorporate heterogeneous (but latent) subpopulations, and a natural goal is to perform well across all of these [57, 65, 21]. While many statistical models show strong average performance, their performance often deteriorates on minority groups underrepresented in the dataset. For example, speech recognition systems are inaccurate for people with minority accents [4]. In numerous other applications--such as facial recognition, automatic video captioning, language identification, academic recommender systems--performance varies significantly over different demographic groupings, such as race, gender, or age [38, 42, 18, 68, 76].

We consider the multi-armed bandit setting with a twist. Rather than having just one decision maker deciding which arm to pull in each round, we have $n$ different decision makers (agents). In the simple stochastic setting we show that one of the agents (called the free rider), who has access to the history of other agents playing some zero regret algorithm can achieve just $O(1)$ regret, as opposed to the regret lower bound of $\Omega (\log T)$ when one decision maker is playing in isolation. In the linear contextual setting, we show that if the other agents play a particular, popular zero regret algorithm (UCB), then the free rider can again achieve $O(1)$ regret. In order to prove this result, we give a deterministic lower bound on the number of times each suboptimal arm must be pulled in UCB. In contrast, we show that the free-rider cannot beat the standard single-player regret bounds in certain partial information settings.