We develop an effective generation of adversarial attacks on neural models that output a sequence of probability distributions rather than a sequence of single values. This setting includes the recently proposed deep probabilistic autoregressive forecasting models that estimate the probability distribution of a time series given its past and achieve state-of-the-art results in a diverse set of application domains. The key technical challenge we address is how to effectively differentiate through the Monte-Carlo estimation of statistics of the output sequence joint distribution. Additionally, we extend prior work on probabilistic forecasting to the Bayesian setting which allows conditioning on future observations, instead of only on past observations. We demonstrate that our approach can successfully generate attacks with small input perturbations in two challenging tasks where robust decision making is crucial – stock market trading and prediction of electricity consumption.

Recent years have seen the rise of statistical program learning based on neural models as an alternative to traditional rule-based systems for programming by example. Rule-based approaches offer correctness guarantees in an unsupervised way as they inherently capture logical rules, while neural models are more realistically scalable to raw, high-dimensional input, and provide resistance to noisy I/O specifications. We introduce PLANS (Program LeArning from Neurally inferred Specifications), a hybrid model for program synthesis from visual observations that gets the best of both worlds, relying on (i) a neural architecture trained to extract abstract, high-level information from each raw individual input (ii) a rule-based system using the extracted information as I/O specifications to synthesize a program capturing the different observations. In order to address the key challenge of making PLANS resistant to noise in the network's output, we introduce a dynamic filtering algorithm for I/O specifications based on selective classification techniques. We obtain state-of-the-art performance at program synthesis from diverse demonstration videos in the Karel and ViZDoom environments, while requiring no ground-truth program for training.

Evolutionary algorithms, being problem-independent and randomized heuristics, are generally believed to be robust to dynamic changes and noisy access to the problem instance. We propose a new method to obtain rigorous runtime results for such settings. In contrast to many previous works, our new approach mostly relies on general parameters of the dynamics or the noise models, such as the expected change of the dynamic optimum or the probability to have a dynamic change in one iteration. Consequently, we obtain bounds which are valid for large varieties of such models. Despite this generality, for almost all particular models regarded in the past our bounds are stronger than those given in previous works. As one particular result, we prove that the (1 + λ) EA can optimize the OneMax benchmark function efficiently despite a constant rate of 1-bit flip noise. For this, a logarithmic size offspring population suffices (the previous-best result required a super-linear value of λ). Our results suggest that the typical way to find the optimum in such adverse settings is not via a steady approach of the optimum, but rather via an exceptionally fast approach after waiting for a rare phase of low dynamic changes or noise.