There is a conjectured computational-statistical gap in terms of the number of samples needed to perform tensor estimation. In particular, for a low rank 3-order tensor with $\Theta(n)$ parameters, Barak and Moitra conjectured that $\Omega(n^{3/2})$ samples are needed for polynomial time computation based on a reduction of a specific hard instance of a rank 1 tensor to the random 3-XOR distinguishability problem. In this paper, we take a complementary perspective and characterize a subclass of tensor instances that can be estimated with only $O(n^{1+\kappa})$ observations for any arbitrarily small constant $\kappa > 0$, nearly linear. If one considers the class of tensors with constant orthogonal CP-rank, the "hardness" of the instance can be parameterized by the minimum absolute value of the sum of latent factor vectors. If the sum of each latent factor vector is bounded away from zero, we present an algorithm that can perform tensor estimation with $O(n^{1+\kappa})$ samples for a $t$-order tensor, significantly less than the previous achievable bound of $O(n^{t/2})$, and close to the lower bound of $\Omega(n)$. This result suggests that amongst constant orthogonal CP-rank tensors, the set of computationally hard instances to estimate are in fact a small subset of all possible tensors.

Deep neural networks are typically trained under a supervised learning framework where a model learns a single task using labeled data. Instead of relying solely on labeled data, practitioners can harness unlabeled or related data to improve model performance, which is often more accessible and ubiquitous. Self-supervised pre-training for transfer learning is becoming an increasingly popular technique to improve state-of-the-art results using unlabeled data. It involves first pre-training a model on a large amount of unlabeled data, then adapting the model to target tasks of interest. In this review, we survey self-supervised learning methods and their applications within the sequential transfer learning framework. We provide an overview of the taxonomy for self-supervised learning and transfer learning, and highlight some prominent methods for designing pre-training tasks across different domains. Finally, we discuss recent trends and suggest areas for future investigation.

While regression tasks aim at interpolating a relation on the entire input space, they often have to be solved with a limited amount of training data. Still, if the hypothesis functions can be sketched well with the data, one can hope for identifying a generalizing model. In this work, we introduce with the Neural Restricted Isometry Property (NeuRIP) a uniform concentration event, in which all shallow $\mathrm{ReLU}$ networks are sketched with the same quality. To derive the sample complexity for achieving NeuRIP, we bound the covering numbers of the networks in the Sub-Gaussian metric and apply chaining techniques. In case of the NeuRIP event, we then provide bounds on the expected risk, which hold for networks in any sublevel set of the empirical risk. We conclude that all networks with sufficiently small empirical risk generalize uniformly.

"Made in China" used to mean cheap and poor quality, and probably involving the theft of intellectual property somewhere along the line. That perception has been out of date for many years now. Counterfeiting by Chinese manufacturers is still a major problem in some industries, but the best Chinese companies are world leaders in quality and in innovation. European telecoms utilities are alarmed by Trump's demand that they exclude Huawei components from their 5G rollout programmes: if they comply, their 5G services will be late and expensive. China's two mobile payments giants, Alibaba's Alipay and Tencent's WeChat Pay, both have many more active users than PayPal and Apple Pay combined.

Mayra Leiva of Reseda, California, knew her eight-year-old son was a little interested in history. But she was surprised when all at once he became a walking encyclopedia, spouting dates and pretending every tire swing was a time machine. "It happened after he saw Night at the Museum," she says. I've had to do a lot of Googling to keep up!" Not many children will tell you that their favorite school subject is history. Memorizing dates and learning long-ago facts that don't seem relevant isn't exactly high on their fun list. Perhaps that's why pop culture--movies, music, television, and even video games and comic books--can be such useful teaching tools. "Teaching through pop culture helps students relate history to their own background and experiences," says Gail Hudson, a fifth-grade teacher and 2020 Nevada Teacher of the Year. "It's tying into something that's already caught their interest." Take the movie version of the Broadway show Hamilton, which releases on Disney July 3.

Black Lives Matter is reverberating around the world, triggering a fresh reckoning with the racist global history of colonialism and slavery. While Confederate statues began to tumble across the American South, in Bristol, England, a diverse group felled a statue of a slave trader that has long provoked offense. Statues of colonial conquerors of Africa and South Asia have followed, along with a robust discussion of the ways in which such actions make history rather than erase it. These movements abroad are not merely echoes of BLM; BLM itself is global. The shared impetus is a common opposition to racism, of which anti-Black racism has been the most lethal and traumatic.

To address the unprecedented scale of HL-LHC data, the Exa.TrkX project is investigating a variety of machine learning approaches to particle track reconstruction. The most promising of these solutions, graph neural networks (GNN), process the event as a graph that connects track measurements (detector hits corresponding to nodes) with candidate line segments between the hits (corresponding to edges). Detector information can be associated with nodes and edges, enabling a GNN to propagate the embedded parameters around the graph and predict node-, edge- and graph-level observables. Previously, message-passing GNNs have shown success in predicting doublet likelihood, and we here report updates on the state-of-the-art architectures for this task. In addition, the Exa.TrkX project has investigated innovations in both graph construction, and embedded representations, in an effort to achieve fully learned end-to-end track finding. Hence, we present a suite of extensions to the original model, with encouraging results for hitgraph classification. In addition, we explore increased performance by constructing graphs from learned representations which contain non-linear metric structure, allowing for efficient clustering and neighborhood queries of data points. We demonstrate how this framework fits in with both traditional clustering pipelines, and GNN approaches. The embedded graphs feed into high-accuracy doublet and triplet classifiers, or can be used as an end-to-end track classifier by clustering in an embedded space. A set of post-processing methods improve performance with knowledge of the detector physics. Finally, we present numerical results on the TrackML particle tracking challenge dataset, where our framework shows favorable results in both seeding and track finding.

Language representations are known to carry stereotypical biases and, as a result, lead to biased predictions in downstream tasks. While existing methods are effective at mitigating biases by linear projection, such methods are too aggressive: they not only remove bias, but also erase valuable information from word embeddings. We develop new measures for evaluating specific information retention that demonstrate the tradeoff between bias removal and information retention. To address this challenge, we propose OSCaR (Orthogonal Subspace Correction and Rectification), a bias-mitigating method that focuses on disentangling biased associations between concepts instead of removing concepts wholesale. Our experiments on gender biases show that OSCaR is a well-balanced approach that ensures that semantic information is retained in the embeddings and bias is also effectively mitigated.

This paper analyses the application of artificial intelligence techniques to various areas of archaeology and more specifically: a) The use of software tools as a creative stimulus for the organization of exhibitions; the use of humanoid robots and holographic displays as guides that interact and involve museum visitors; b) The analysis of methods for the classification of fragments found in archaeological excavations and for the reconstruction of ceramics, with the recomposition of the parts of text missing from historical documents and epigraphs; c) The cataloguing and study of human remains to understand the social and historical context of belonging with the demonstration of the effectiveness of the AI techniques used; d) The detection of particularly difficult terrestrial archaeological sites with the analysis of the architectures of the Artificial Neural Networks most suitable for solving the problems presented by the site; the design of a study for the exploration of marine archaeological sites, located at depths that cannot be reached by man, through the construction of a freely explorable 3D version.

Deep Reinforcement Learning has proved to be able to solve many control tasks in different fields, but the behavior of these systems is not always as expected when deployed in real-world scenarios. This is mainly due to the lack of domain adaptation between simulated and real-world data together with the absence of distinction between train and test datasets. In this work, we investigate these problems in the autonomous driving field, especially for a maneuver planning module for roundabout insertions. In particular, we present a system based on multiple environments in which agents are trained simultaneously, evaluating the behavior of the model in different scenarios. Finally, we analyze techniques aimed at reducing the gap between simulated and real-world data showing that this increased the generalization capabilities of the system both on unseen and real-world scenarios.