In singular models, the optimal set of parameters forms an analytic set with singularities and classical statistical inference cannot be applied to such models. This is significant for deep learning as neural networks are singular and thus "dividing" by the determinant of the Hessian or employing the Laplace approximation are not appropriate. Despite its potential for addressing fundamental issues in deep learning, singular learning theory appears to have made little inroads into the developing canon of deep learning theory. Via a mix of theory and experiment, we present an invitation to singular learning theory as a vehicle for understanding deep learning and suggest important future work to make singular learning theory directly applicable to how deep learning is performed in practice.

Submodular functions allow to model many real-world optimisation problems. This paper introduces approaches for computing diverse sets of high quality solutions for submodular optimisation problems. We first present diversifying greedy sampling approaches and analyse them with respect to the diversity measured by entropy and the approximation quality of the obtained solutions. Afterwards, we introduce an evolutionary diversity optimisation approach to further improve diversity of the set of solutions. We carry out experimental investigations on popular submodular benchmark functions that show that the combined approaches achieve high quality solutions of large diversity.

Personalized recommendation based on multi-arm bandit (MAB) algorithms has shown to lead to high utility and efficiency as it can dynamically adapt the recommendation strategy based on feedback. However, unfairness could incur in personalized recommendation. In this paper, we study how to achieve user-side fairness in personalized recommendation. We formulate our fair personalized recommendation as a modified contextual bandit and focus on achieving fairness on the individual whom is being recommended an item as opposed to achieving fairness on the items that are being recommended. We introduce and define a metric that captures the fairness in terms of rewards received for both the privileged and protected groups. We develop a fair contextual bandit algorithm, Fair-LinUCB, that improves upon the traditional LinUCB algorithm to achieve group-level fairness of users. Our algorithm detects and monitors unfairness while it learns to recommend personalized videos to students to achieve high efficiency. We provide a theoretical regret analysis and show that our algorithm has a slightly higher regret bound than LinUCB. We conduct numerous experimental evaluations to compare the performances of our fair contextual bandit to that of LinUCB and show that our approach achieves group-level fairness while maintaining a high utility.

Directional data consist of observations distributed on a (hyper)sphere, and appear in many applied fields, such as astronomy, ecology, and environmental science. This paper studies both statistical and computational problems of kernel smoothing for directional data. We generalize the classical mean shift algorithm to directional data, which allows us to identify local modes of the directional kernel density estimator (KDE). The statistical convergence rates of the directional KDE and its derivatives are derived, and the problem of mode estimation is examined. We also prove the ascending property of our directional mean shift algorithm and investigate a general problem of gradient ascent on the unit hypersphere. To demonstrate the applicability of our proposed algorithm, we evaluate it as a mode clustering method on both simulated and real-world datasets.

Deep learning classifiers are susceptible to well-crafted, imperceptible variations of their inputs, known as adversarial attacks. In this regard, the study of powerful attack models sheds light on the sources of vulnerability in these classifiers, hopefully leading to more robust ones. In this paper, we introduce AdvFlow: a novel black-box adversarial attack method on image classifiers that exploits the power of normalizing flows to model the density of adversarial examples around a given target image. We see that the proposed method generates adversaries that closely follow the clean data distribution, a property which makes their detection less likely. Also, our experimental results show competitive performance of the proposed approach with some of the existing attack methods on defended classifiers. The code is available at https://github.com/hmdolatabadi/AdvFlow.

In recent years, deep neural networks have been successful in both industry and academia, especially for computer vision tasks. The great success of deep learning is mainly due to its scalability to encode large-scale data and to maneuver billions of model parameters. However, it is a challenge to deploy these cumbersome deep models on devices with limited resources, e.g., mobile phones and embedded devices, not only because of the high computational complexity but also the large storage requirements. To this end, a variety of model compression and acceleration techniques have been developed. As a representative type of model compression and acceleration, knowledge distillation effectively learns a small student model from a large teacher model. It has received rapid increasing attention from the community. This paper provides a comprehensive survey of knowledge distillation from the perspectives of knowledge categories, training schemes, teacher-student architecture, distillation algorithms, performance comparison and applications. Furthermore, challenges in knowledge distillation are briefly reviewed and comments on future research are discussed and forwarded.

We present a framework to derive bounds on the test loss of randomized learning algorithms for the case of bounded loss functions. This framework leads to bounds that depend on the conditional information density between the the output hypothesis and the choice of the training set, given a larger set of data samples from which the training set is formed. If the conditional information density is bounded uniformly in the sizenof the training set, our bounds decay as1/n, which is referred to as a fast rate. This is in contrast with the tail bounds involving conditional information measures available in the literature, which have a less benign 1/ n dependence. We demonstrate the usefulness of our tail bounds by showing that they lead to estimates of the test loss achievable with several neural network architectures trained on MNIST and Fashion-MNIST that match the state-of-the-art bounds available in the literature. In recent years, there has been a surge of interest in the use of information-theoretic techniques for bounding the loss of learning algorithms. While the first results of this flavor can be traced to the probably approximately correct (PAC)-Bayesian approach (McAllester, 1998; Catoni, 2007) (see also (Guedj, 2019) for a recent review), the connection between loss bounds and classical information-theoretic measures was made explicit in the works of Russo & Zou (2016) and Xu & Raginsky (2017), where bounds on the average population loss were derived in terms of the mutual information between the training data and the output hypothesis. Since then, these average loss bounds have been tightened (Bu et al., 2019; Asadi et al., 2018; Negrea et al., 2019). Furthermore, the information-theoretic framework has also been successfully applied to derive tail probability bounds on the population loss (Bassily et al., 2018; Esposito et al., 2019; Hellström & Durisi, 2020a). Of particular relevance to the present paper is the random-subset setting, introduced by Steinke & Zakynthinou (2020) and further studied in (Hellström & Durisi, 2020b; Haghifam et al., 2020).

The task of learning a causal model from observational data, or a combination of observational and interventional data, is commonly referred to as a causal discovery or causal structure learning [1]. Causal discovery from two variables based on observational data in the absence of time series or controlled interventions is a challenging problem and necessitates additional assumptions [2]. This is a ubiquitous problem in almost all domains of science, but particularly so in econometrics, meteorology, biology and medicine where interventional approaches are difficult or in several cases not feasible. Model-free data-driven approaches for causal discovery have developed significantly over the past decade or so in an attempt to address the problem of causal discovery such as Granger Causality (GC) [3], Transfer Entropy (TE) [4] and Compression-Complexity Causality (CCC) [5]. These methods have been used in various disciplines across neuroscience, climatology, econometrics, etc and rely on properties of time-series data. Both GC and TE have assumptions that need to be met for satisfactory inference, while CCC is assumption-free and robust to many artefacts and nuisance variables. All three need careful parameter calibration and selection for optimally accurate performance. A class of model-free causal discovery methods do not assume a temporal structure in the data and are rooted in algorithmic information theory, chiefly based on the notion of Kolmogorov complexity. The Kolmogorov complexity of a finite binary string is the length of the shortest binary program that generates that string and reflects the computational resources needed to specify it.

Large scale recommender models find most relevant items from huge catalogs, and they play a critical role in modern search and recommendation systems. To model the input space with large-vocab categorical features, a typical recommender model learns a joint embedding space through neural networks for both queries and items from user feedback data. However, with millions to billions of items, the power-law user feedback makes labels very sparse for a large amount of long-tail items. Inspired by the recent success in self-supervised representation learning research in both computer vision and natural language understanding, we propose a multi-task self-supervised learning (SSL) framework for large-scale item recommendations. The framework is designed to tackle the label sparsity problem by learning more robust item representations. Furthermore, we propose two self-supervised tasks applicable to models with categorical features within the proposed framework: (i) Feature Masking (FM) and (ii) Feature Dropout (FD). We evaluate our framework using two large-scale datasets with 500M and 1B training examples respectively. Our results demonstrate that the proposed framework outperforms traditional supervised learning only models and state-of-the-art regularization techniques in the context of item recommendations. The SSL framework shows larger improvement with less supervision compared to the counterparts. We also apply the proposed techniques to a web-scale commercial app-to-app recommendation system, and significantly improve top-tier business metrics via A/B experiments on live traffic. Our online results also verify our hypothesis that our framework indeed improves model performance on slices that lack supervision.

Coresets are small data summaries that are sufficient for model training. They can be maintained online, enabling efficient handling of large data streams under resource constraints. However, existing constructions are limited to simple models such as k-means and logistic regression. In this work, we propose a novel coreset construction via cardinality-constrained bilevel optimization. We show how our framework can efficiently generate coresets for deep neural networks, and demonstrate its empirical benefits in continual learning and in streaming settings.