We propose transferability from Large Geometric Vicinity (LGV), a new technique to increase the transferability of black-box adversarial attacks. LGV starts from a pretrained surrogate model and collects multiple weight sets from a few additional training epochs with a constant and high learning rate. LGV exploits two geometric properties that we relate to transferability. First, models that belong to a wider weight optimum are better surrogates. Second, we identify a subspace able to generate an effective surrogate ensemble among this wider optimum. Through extensive experiments, we show that LGV alone outperforms all (combinations of) four established test-time transformations by 1.8 to 59.9 percentage points. Our findings shed new light on the importance of the geometry of the weight space to explain the transferability of adversarial examples.

Unsupervised semantic segmentation requires assigning a label to every pixel without any human annotations. Despite recent advances in self-supervised representation learning for individual images, unsupervised semantic segmentation with pixel-level representations is still a challenging task and remains underexplored. In this work, we propose a self-supervised pixel representation learning method for semantic segmentation by using visual concepts (i.e., groups of pixels with semantic meanings, such as parts, objects, and scenes) extracted from images. To guide self-supervised learning, we leverage three types of relationships between pixels and concepts, including the relationships between pixels and local concepts, local and global concepts, as well as the co-occurrence of concepts. We evaluate the learned pixel embeddings and visual concepts on three datasets, including PASCAL VOC 2012, COCO 2017, and DAVIS 2017. Our results show that the proposed method gains consistent and substantial improvements over recent unsupervised semantic segmentation approaches, and also demonstrate that visual concepts can reveal insights into image datasets.

Decision trees are highly famous in machine learning and usually acquire state-of-the-art performance. Despite that, well-known variants like CART, ID3, random forest, and boosted trees miss a probabilistic version that encodes prior assumptions about tree structures and shares statistical strength between node parameters. Existing work on Bayesian decision trees depend on Markov Chain Monte Carlo (MCMC), which can be computationally slow, especially on high dimensional data and expensive proposals. In this study, we propose a method to parallelise a single MCMC decision tree chain on an average laptop or personal computer that enables us to reduce its run-time through multi-core processing while the results are statistically identical to conventional sequential implementation. We also calculate the theoretical and practical reduction in run time, which can be obtained utilising our method on multi-processor architectures. Experiments showed that we could achieve 18 times faster running time provided that the serial and the parallel implementation are statistically identical.

Learning a parametric model of a data distribution is a well-known statistical problem that has seen renewed interest as it is brought to scale in deep learning. Framing the problem as a self-supervised task, where data samples are discriminated from noise samples, is at the core of state-of-the-art methods, beginning with Noise-Contrastive Estimation (NCE). Yet, such contrastive learning requires a good noise distribution, which is hard to specify; domain-specific heuristics are therefore widely used. While a comprehensive theory is missing, it is widely assumed that the optimal noise should in practice be made equal to the data, both in distribution and proportion. This setting underlies Generative Adversarial Networks (GANs) in particular. Here, we empirically and theoretically challenge this assumption on the optimal noise. We show that deviating from this assumption can actually lead to better statistical estimators, in terms of asymptotic variance. In particular, the optimal noise distribution is different from the data's and even from a different family.

Linear algebra is a mathematical discipline concerned with studying vector spaces and linear mappings between them [1]. It is essential in artificial intelligence implementations because it allows for unlocking meanings in high-dimensional data, a common use case pipeline (in AI). Applications for implementation include solving problems across many use cases in AI, including machine learning, deep learning, and natural language processing. Namely, it can be utilized to predict the behavior of neural networks, and it is also being used to improve the accuracy of deep learning models. Further, linear algebra provides a way to understand and visualize high-dimensional data, often used in natural language processing tasks.

This article was published as a part of the Data Science Blogathon. In this article, we will explore different types of gradient descent. So let's get started with the article. The algorithm designer can set the learning rate. If we use a learning rate that is too small, it will cause us to update very slowly, requiring more iterations to get a better solution.

One of the key issues with drug discovery is dealing with uncertainty on outcome timing, having few observations in your patient samples, and being overwhelmed with features. In this article I discuss some of the data science methods used previously and new approaches my team and I have been able to take resulting in success. Using AI for drug discovery is one of the most exciting domains that is already starting to impact and improve our society. Suppose you are asked to compare proportions-risks, rates, etc… between different groups. If you are in the realm of statistical testing, you will be using a chi-square or the Fisher-Exact test, or Logistic Regression and allied models beyond that.

What is a Generalized Linear Model? Why GLM? Assumptions of GLM Components of GLM Different Generalized Linear Models Difference Between Generalized Linear Model and General Linear Model Can Generalized Linear Models have correlated data? Generalized Linear Model (GLiM, or GLM) is an advanced statistical modelling technique formulated by John Nelder and Robert Wedderburn in 1972. It is an umbrella term that encompasses many other models, which allows the response variable y to have an error distribution other than a normal distribution. The models include Linear Regression, Logistic Regression, and Poisson Regression.

A hallmark of human intelligence is the ability to construct self-contained chunks of knowledge and reuse them in novel combinations for solving different problems. Learning such compositional structures has been a challenge for artificial systems, due to the underlying combinatorial search. To date, research into compositional learning has largely proceeded separately from work on lifelong or continual learning. This dissertation integrated these two lines of work to present a general-purpose framework for lifelong learning of functionally compositional structures. The framework separates the learning into two stages: learning how to combine existing components to assimilate a novel problem, and learning how to adapt the existing components to accommodate the new problem. This separation explicitly handles the trade-off between stability and flexibility. This dissertation instantiated the framework into various supervised and reinforcement learning (RL) algorithms. Supervised learning evaluations found that 1) compositional models improve lifelong learning of diverse tasks, 2) the multi-stage process permits lifelong learning of compositional knowledge, and 3) the components learned by the framework represent self-contained and reusable functions. Similar RL evaluations demonstrated that 1) algorithms under the framework accelerate the discovery of high-performing policies, and 2) these algorithms retain or improve performance on previously learned tasks. The dissertation extended one lifelong compositional RL algorithm to the nonstationary setting, where the task distribution varies over time, and found that modularity permits individually tracking changes to different elements in the environment. The final contribution of this dissertation was a new benchmark for compositional RL, which exposed that existing methods struggle to discover the compositional properties of the environment.

In this chapter, we identify fundamental geometric structures that underlie the problems of sampling, optimisation, inference and adaptive decision-making. Based on this identification, we derive algorithms that exploit these geometric structures to solve these problems efficiently. We show that a wide range of geometric theories emerge naturally in these fields, ranging from measure-preserving processes, information divergences, Poisson geometry, and geometric integration. Specifically, we explain how (i) leveraging the symplectic geometry of Hamiltonian systems enable us to construct (accelerated) sampling and optimisation methods, (ii) the theory of Hilbertian subspaces and Stein operators provides a general methodology to obtain robust estimators, (iii) preserving the information geometry of decision-making yields adaptive agents that perform active inference. Throughout, we emphasise the rich connections between these fields; e.g., inference draws on sampling and optimisation, and adaptive decision-making assesses decisions by inferring their counterfactual consequences. Our exposition provides a conceptual overview of underlying ideas, rather than a technical discussion, which can be found in the references herein.