Many modern data sets are sampled with error from complex high-dimensional surfaces. Methods such as tensor product splines or Gaussian processes are effective/well suited for characterizing a surface in two or three dimensions but may suffer from difficulties when representing higher dimensional surfaces. Motivated by high throughput toxicity testing where observed dose-response curves are cross sections of a surface defined by a chemical's structural properties, a model is developed to characterize this surface to predict untested chemicals' dose-responses. This manuscript proposes a novel approach that models the multidimensional surface as a sum of learned basis functions formed as the tensor product of lower dimensional functions, which are themselves representable by a basis expansion learned from the data. The model is described, a Gibbs sampling algorithm proposed, and is investigated in a simulation study as well as data taken from the US EPA's ToxCast high throughput toxicity testing platform.

An implementation of a nonparametric Bayesian approach to solving binary classification problems on graphs is described. A hierarchical Bayesian approach with a randomly scaled Gaussian prior is considered. The prior uses the graph Laplacian to take into account the underlying geometry of the graph. A method based on a theoretically optimal prior and a more flexible variant using partial conjugacy are proposed. Two simulated data examples and two examples using real data are used in order to illustrate the proposed methods.

Consider the problem of estimating click probabilities for links between pages of a website, given a hyperlink graph and aggregate statistics on the number of times each page has been visited. Naively, one might expect that the probability of clicking on a particular link should be roughly proportional to the traffic of the link's target. However, this neglects important structural effects: a page's traffic is influenced by a) the number of incoming links, b) the traffic at the pages that link to it, and c) the traffic absorbed by competing links. In order to successfully infer click probabilities, it is therefore necessary to disentangle the preference for a page (i.e., the intrinsic propensity of a user to click on a link pointing to it) from the page's visibility (the exposure it gets from pages linking to it). Building upon recent work by Kumar et al. [2015], we present a statistical framework that tackles a general formulation of the problem: given a network (representing possible transitions between nodes) and the marginal traffic at each node, recover the transition probabilities.

We address the problem of learning a ranking by using adaptively chosen pairwise comparisons. Our goal is to recover the ranking accurately but to sample the comparisons sparingly. If all comparison outcomes are consistent with the ranking, the optimal solution is to use an efficient sorting algorithm, such as Quicksort. But how do sorting algorithms behave if some comparison outcomes are inconsistent with the ranking? We give favorable guarantees for Quicksort for the popular Bradley-Terry model, under natural assumptions on the parameters. Furthermore, we empirically demonstrate that sorting algorithms lead to a very simple and effective active learning strategy: repeatedly sort the items. This strategy performs as well as state-of-the-art methods (and much better than random sampling) at a minuscule fraction of the computational cost.

Deep clustering is the first method to handle general audio separation scenarios with multiple sources of the same type and an arbitrary number of sources, performing impressively in speaker-independent speech separation tasks. However, little is known about its effectiveness in other challenging situations such as music source separation. Contrary to conventional networks that directly estimate the source signals, deep clustering generates an embedding for each time-frequency bin, and separates sources by clustering the bins in the embedding space. We show that deep clustering outperforms conventional networks on a singing voice separation task, in both matched and mismatched conditions, even though conventional networks have the advantage of end-to-end training for best signal approximation, presumably because its more flexible objective engenders better regularization. Since the strengths of deep clustering and conventional network architectures appear complementary, we explore combining them in a single hybrid network trained via an approach akin to multi-task learning. Remarkably, the combination significantly outperforms either of its components.

Kohn-Sham density functional theory[1] is now enormously popular as an electronic structure method in a wide variety of fields[2]. Useful accuracy is achieved with standard exchange-correlation approximations, such as generalized gradient approximations[3] and hybrids[4]. Such calculations are playing a key role in the materials genome initiative[5], at least for weakly correlated materials[6]. There has also been a recent spike of interest in applying machine learning (ML) methods in the physical sciences[7-11]. The majority of these applications involve predicting properties of molecules or materials from large databases of KS-DFT calculations[12-15]. A few applications involve finding potential energy surfaces within MD simulations[16-19]. Fewer still have focussed on finding the functionals of DFT as a method of performing KS electronic structure calculations without solving the KS equations[20-23]. If such attempts could be made practical, the possible speedup in repeated DFT calculations of similar species, such as occur in ab initio MD simulations, is enormous. A key difficulty has been the need to extract the functional derivative of the non-interacting kinetic energy.

We introduce a simple and accurate neural model for dependency-based semantic role labeling. Our model predicts predicate-argument dependencies relying on states of a bidirectional LSTM encoder. The semantic role labeler achieves competitive performance on English, even without any kind of syntactic information and only using local inference. However, when automatically predicted part-of-speech tags are provided as input, it substantially outperforms all previous local models and approaches the best reported results on the English CoNLL-2009 dataset. We also consider Chinese, Czech and Spanish where our approach also achieves competitive results. Syntactic parsers are unreliable on out-of-domain data, so standard (i.e., syntactically-informed) SRL models are hindered when tested in this setting. Our syntax-agnostic model appears more robust, resulting in the best reported results on standard out-of-domain test sets.

It is clear that one of the primary tools we can use to mitigate the potential risk from a misbehaving AI system is the ability to turn the system off. As the capabilities of AI systems improve, it is important to ensure that such systems do not adopt subgoals that prevent a human from switching them off. This is a challenge because many formulations of rational agents create strong incentives for self-preservation. This is not caused by a built-in instinct, but because a rational agent will maximize expected utility and cannot achieve whatever objective it has been given if it is dead. Our goal is to study the incentives an agent has to allow itself to be switched off. We analyze a simple game between a human H and a robot R, where H can press R's off switch but R can disable the off switch. A traditional agent takes its reward function for granted: we show that such agents have an incentive to disable the off switch, except in the special case where H is perfectly rational. Our key insight is that for R to want to preserve its off switch, it needs to be uncertain about the utility associated with the outcome, and to treat H's actions as important observations about that utility. (R also has no incentive to switch itself off in this setting.) We conclude that giving machines an appropriate level of uncertainty about their objectives leads to safer designs, and we argue that this setting is a useful generalization of the classical AI paradigm of rational agents.

U.S. officials reportedly are rethinking the advisability of allowing the Chinese to invest in sensitive technologies seen as vital to national security. Reuters reported Wednesday U.S. officials are concerned such cutting-edge technologies as artificial intelligence and machine learning could be used by the Chinese to augment their military capabilities and achieve greater advancements in strategic industries. Technology is the fastest growing industry in the United States, and China has funneled $45.6 billion into U.S. acquisitions and Greenfield investments in the last year, Rhodium Group found. That investment is expected to double this year. Read: What Is Artificial Intelligence?

Montreal-based Element AI, a key player in the city's burgeoning artificial-intelligence sector, has clinched a major financing deal to fund future growth and job creation. Element is set to announce on Wednesday that it has raised US$102-million from a group of investors led by San Francisco venture capital fund Data Collective (DCVC). The deal is the largest Series A funding round for an AI company in history, Element says. The investment will allow Element to "accelerate its capabilities and invest in large-scale AI projects internationally, solidifying its position as the largest global AI company in Canada and creating 250 jobs in the Canadian high tech sector by January 2018," it said in a news release. Element was founded last year by tech entrepreneurs Jean-François Gagné and Nicolas Chapados, Montreal venture capital fund Real Ventures, and Université de Montréal AI scientist Yoshua Bengio.