Neural networks are vulnerable to adversarial examples, i.e. inputs that are imperceptibly perturbed from natural data and yet incorrectly classified by the network. Adversarial training, a heuristic form of robust optimization that alternates between minimization and maximization steps, has proven to be among the most successful methods to train networks that are robust against a pre-defined family of perturbations. This paper provides a partial answer to the success of adversarial training. When the inner maximization problem can be solved to optimality, we prove that adversarial training finds a network of small robust train loss. When the maximization problem is solved by a heuristic algorithm, we prove that adversarial training finds a network of small robust surrogate train loss. The analysis technique leverages recent work on the analysis of neural networks via Neural Tangent Kernel (NTK), combined with online-learning when the maximization is solved by a heuristic, and the expressiveness of the NTK kernel in the $\ell_\infty$-norm.

We study three fundamental statistical-learning problems: distribution estimation, property estimation, and property testing. We establish the profile maximum likelihood (PML) estimator as the first unified sample-optimal approach to a wide range of learning tasks. In particular, for every alphabet size $k$ and desired accuracy $\varepsilon$: $\textbf{Distribution estimation}$ Under $\ell_1$ distance, PML yields optimal $\Theta(k/(\varepsilon^2\log k))$ sample complexity for sorted-distribution estimation, and a PML-based estimator empirically outperforms the Good-Turing estimator on the actual distribution; $\textbf{Additive property estimation}$ For a broad class of additive properties, the PML plug-in estimator uses just four times the sample size required by the best estimator to achieve roughly twice its error, with exponentially higher confidence; $\boldsymbol{\alpha}\textbf{-R\'enyi entropy estimation}$ For integer $\alpha>1$, the PML plug-in estimator has optimal $k^{1-1/\alpha}$ sample complexity; for non-integer $\alpha>3/4$, the PML plug-in estimator has sample complexity lower than the state of the art; $\textbf{Identity testing}$ In testing whether an unknown distribution is equal to or at least $\varepsilon$ far from a given distribution in $\ell_1$ distance, a PML-based tester achieves the optimal sample complexity up to logarithmic factors of $k$. With minor modifications, most of these results also hold for a near-linear-time computable variant of PML.

We present a novel extension to embedding-based knowledge graph completion models which enables them to perform open-world link prediction, i.e. to predict facts for entities unseen in training based on their textual description. Our model combines a regular link prediction model learned from a knowledge graph with word embeddings learned from a textual corpus. After training both independently, we learn a transformation to map the embeddings of an entity's name and description to the graph-based embedding space. In experiments on several datasets including FB20k, DBPedia50k and our new dataset FB15k-237-OWE, we demonstrate competitive results. Particularly, our approach exploits the full knowledge graph structure even when textual descriptions are scarce, does not require a joint training on graph and text, and can be applied to any embedding-based link prediction model, such as TransE, ComplEx and DistMult.

We consider a problem of manifold estimation from noisy observations. We suggest a novel adaptive procedure, which simultaneously reconstructs a smooth manifold from the observations and estimates projectors onto the tangent spaces. Many manifold learning procedures locally approximate a manifold by a weighted average over a small neighborhood. However, in the presence of large noise, the assigned weights become so corrupted that the averaged estimate shows very poor performance. We adjust the weights so they capture the manifold structure better. We propose a computationally efficient procedure, which iteratively refines the weights on each step, such that, after several iterations, we obtain the "oracle" weights, so the quality of the final estimates does not suffer even in the presence of relatively large noise. We also provide a theoretical study of the procedure and prove its optimality deriving both new upper and lower bounds for manifold estimation under the Hausdorff loss.

There are exciting opportunities if you're thinking about higher education, reskilling or specialising. Government has introduced a new industry funded AI Masters programme, beginning with at least 200 new AI Masters students in September 2019. As part of the Industrial Strategy, Government is committed to putting the UK at the forefront of the artificial intelligence and data revolution. A Masters degree is a relatively quick way to upskill existing employees, returners to work, or individuals interested in converting from other disciplines. It is expected that this programme will expand to include more students year-on-year.

Imagine a human-like teacher with no human flaws. The best educators in the world sometimes suffer from innate human errors, taking different forms in every one of us. They will eventually grow tired and nervous. Not even the best of them can provide personal attention to a class of 30. Computers never sleep; the knowledge they impart is available 24/7 across continents, time zones, and devices.

It seems like everyone wants to be a data scientist these days -- from PhD students to data analysts to your old college roommate who keeps Linkedin messaging you to'grab coffee'. Perhaps you've had the same inkling that you should at least explore some data science positions and see what the hype is about. Maybe you've seen articles like Vicki Boykis' Data Science is different now that states: Concepts like unit testing and continuous integration rapidly found its way into the jargon and the toolset commonly used by data scientist and numerical scientist working on ML engineering. What's not clear is how you can leverage your experience as a software engineer into a data science position. Are there best practices or tools that are different for data scientists?

Imitation learning is the process by which one agent tries to learn how to perform a certain task using information generated by another, often more-expert agent performing that same task. Conventionally, the imitator has access to both state and action information generated by an expert performing the task (e.g., the expert may provide a kinesthetic demonstration of object placement using a robotic arm). However, requiring the action information prevents imitation learning from a large number of existing valuable learning resources such as online videos of humans performing tasks. To overcome this issue, the specific problem of imitation from observation (IfO) has recently garnered a great deal of attention, in which the imitator only has access to the state information (e.g., video frames) generated by the expert. In this paper, we provide a literature review of methods developed for IfO, and then point out some open research problems and potential future work.

Learning about many things can provide numerous benefits to a reinforcement learning system. For example, learning many auxiliary value functions, in addition to optimizing the environmental reward, appears to improve both exploration and representation learning. The question we tackle in this paper is how to sculpt the stream of experience---how to adapt the system's behaviour---to optimize the learning of a collection of value functions. A simple answer is to compute an intrinsic reward based on the statistics of each auxiliary learner, and use reinforcement learning to maximize that intrinsic reward. Unfortunately, implementing this simple idea has proven difficult, and thus has been the focus of decades of study. It remains unclear which of the many possible measures of learning would work well in a parallel learning setting where environmental reward is extremely sparse or absent. In this paper, we investigate and compare different intrinsic reward mechanisms in a new bandit-like parallel-learning testbed. We discuss the interaction between reward and prediction learners and highlight the importance of introspective prediction learners: those that increase their rate of learning when progress is possible, and decrease when it is not. We provide a comprehensive empirical comparison of 15 different rewards, including well-known ideas from reinforcement learning and active learning. Our results highlight a simple but seemingly powerful principle: intrinsic rewards based on the amount of learning can generate useful behaviour, if each individual learner is introspective.

Data-driven approaches are becoming more common as problem-solving techniques in many areas of research and industry. In most cases, machine learning models are the key component of these solutions, but a solution involves multiple such models, along with significant levels of reasoning with the models' output and input. Current technologies do not make such techniques easy to use for application experts who are not fluent in machine learning nor for machine learning experts who aim at testing ideas and models on real-world data in the context of the overall AI system. We review key efforts made by various AI communities to provide languages for high-level abstractions over learning and reasoning techniques needed for designing complex AI systems. We classify the existing frameworks based on the type of techniques and the data and knowledge representations they use, provide a comparative study of the way they address the challenges of programming real-world applications, and highlight some shortcomings and future directions.