The size of the datasets available today leads to distribute Machine Learning (ML) tasks. An SGD--based optimization is for instance typically carried out by two categories of participants: parameter servers and workers. Some of these nodes can sometimes behave arbitrarily (called \emph{Byzantine} and caused by corrupt/bogus data/machines), impacting the accuracy of the entire learning activity. Several approaches recently studied how to tolerate Byzantine workers, while assuming honest and trusted parameter servers. In order to achieve total ML robustness, we introduce GuanYu, the first algorithm (to the best of our knowledge) to handle Byzantine parameter servers as well as Byzantine workers. We prove that GuanYu ensures convergence against $\frac{1}{3}$ Byzantine parameter servers and $\frac{1}{3}$ Byzantine workers, which is optimal in asynchronous networks (GuanYu does also tolerate unbounded communication delays, i.e.\ asynchrony). To prove the Byzantine resilience of GuanYu, we use a contraction argument, leveraging geometric properties of the median in high dimensional spaces to prevent (with probability 1) any drift on the models within each of the non-Byzantine servers. % To convey its practicality, we implemented GuanYu using the low-level TensorFlow APIs and deployed it in a distributed setup using the CIFAR-10 dataset. The overhead of tolerating Byzantine participants, compared to a vanilla TensorFlow deployment that is vulnerable to a single Byzantine participant, is around 30\% in terms of throughput (model updates per second) - while maintaining the same convergence rate (model updates required to reach some accuracy).

For network architecture search (NAS), it is crucial but challenging to simultaneously guarantee both effectiveness and efficiency. Towards achieving this goal, we develop a differentiable NAS solution, where the search space includes arbitrary feed-forward network consisting of the predefined number of connections. Benefiting from a proposed ensemble Gumbel-Softmax estimator, our method optimizes both the architecture of a deep network and its parameters in the same round of backward propagation, yielding an end-to-end mechanism of searching network architectures. Extensive experiments on a variety of popular datasets strongly evidence that our method is capable of discovering high-performance architectures, while guaranteeing the requisite efficiency during searching.

We describe a method for unmixing mixtures of freely independent random variables in a manner analogous to the independent component analysis (ICA) based method for unmixing independent random variables from their additive mixtures. Random matrices play the role of free random variables in this context so the method we develop, which we call Free component analysis (FCA), unmixes matrices from additive mixtures of matrices. We describe the theory, the various algorithms, and compare FCA to ICA. We show that FCA performs comparably to, and often better than, ICA in every application, such as image and speech unmixing, where ICA has been known to succeed. Our computational experiments suggest that not-so-random matrices, such as images and spectrograms of waveforms are (closer to being) freer "in the wild" than we might have theoretically expected.

Traditional clustering methods often perform clustering with low-level indiscriminative representations and ignore relationships between patterns, resulting in slight achievements in the era of deep learning. To handle this problem, we develop Deep Discriminative Clustering (DDC) that models the clustering task by investigating relationships between patterns with a deep neural network. Technically, a global constraint is introduced to adaptively estimate the relationships, and a local constraint is developed to endow the network with the capability of learning high-level discriminative representations. By iteratively training the network and estimating the relationships in a mini-batch manner, DDC theoretically converges and the trained network enables to generate a group of discriminative representations that can be treated as clustering centers for straightway clustering. Extensive experiments strongly demonstrate that DDC outperforms current methods on eight image, text and audio datasets concurrently.

If there was something that always frustrated me was not fully understanding Bayesian inference. Sometime last year, I came across an article about a TensorFlow-supported R package for Bayesian analysis, called greta. Back then, I searched for greta tutorials and stumbled on this blog post that praised a textbook called Statistical Rethinking: A Bayesian Course with Examples in R and Stan by Richard McElreath. I had found a solution to my lingering frustration so I bought a copy straight away. I spent the last few months reading it cover to cover and solving the proposed exercises, which are heavily based on the rethinking package. I cannot recommend it highly enough to whoever seeks a solid grip on Bayesian statistics, both in theory and application. This post ought to be my most gratifying blogging experience so far, in that I am essentially reporting my own recent learning. I am convinced this will make the storytelling all the more effective. As a demonstration, the female cuckoo reproductive output data recently analysed by Riehl et al., 2019 [1] will be modelled using In the process, we will conduct the MCMC sampling, visualise posterior distributions, generate predictions and ultimately assess the influence of social parasitism in female reproductive output. You should have some familiarity with standard statistical models. If you need to refresh some basics of probabilities using R have a look into my first post. I hope you enjoy as much as I did!

The increasingly digital economy requires boards and executives to have a solid understanding of the rapidly changing digital landscape. Naturally, artificial intelligence (AI) is an important stakeholder. Those organisations that want to prepare for an automated future should have a thorough understanding of AI. However, AI is an umbrella term that covers multiple disciplines, each affecting the business in a slightly different way. Artificial intelligence consists of the seamless integration of robotics, cognitive systems and machine learning.

Would people who are strong in math be good in machine learning? Certainly having a strong background in mathematics will make it easier to understand machine learning at a conceptual level. When someone introduces you to the inference function in logistic regression, you'll say, "Hey, that's just linear algebra!" But surely deep learning must be something new? Not harder, just more (thank God for automatic differentiation). Supposing you did well in those undergrad classes, machine learning should come easy enough at first.

Models often need to be constrained to a certain size for them to be considered interpretable, for e.g., a decision tree of depth 5 is much easier to make sense of than one of depth 30. This suggests a trade-off between interpretability and accuracy. Our work tries to minimize this trade-off by suggesting the optimal distribution of the data to learn from, that surprisingly, may be different from the original distribution. We use an Infinite Beta Mixture Model (IBMM) to represent a specific set of sampling schemes. The parameters of the IBMM are learned using a Bayesian Optimizer (BO). While even under simplistic assumptions a distribution in the original $d$-dimensional space would need to optimize for $O(d)$ variables - cumbersome for most real-world data - our technique lowers this number significantly to a fixed set of 8 variables at the cost of some additional preprocessing. The proposed technique is \emph{model-agnostic}; it can be applied to any classifier. It also admits a general notion of model size. We demonstrate its effectiveness using multiple real-world datasets to construct decision trees, linear probability models and gradient boosted models.

We study the robustness to symmetric label noise of GNNs training procedures. By combining the nonlinear neural message-passing models (e.g. Graph Isomorphism Networks, GraphSAGE, etc.) with loss correction methods, we present a noise-tolerant approach for the graph classification task. Our experiments show that test accuracy can be improved under the artificial symmetric noisy setting.

We present a technique to perform dimensionality reduction on data that is subject to uncertainty. Our method is a generalization of traditional principal component analysis (PCA) to multivariate probability distributions. In comparison to non-linear methods, linear dimensionality reduction techniques have the advantage that the characteristics of such probability distributions remain intact after projection. We derive a representation of the covariance matrix that respects potential uncertainty in each of the observations, building the mathematical foundation of our new method uncertainty-aware PCA. In addition to the accuracy and performance gained by our approach over sampling-based strategies, our formulation allows us to perform sensitivity analysis with regard to the uncertainty in the data. For this, we propose factor traces as a novel visualization that enables us to better understand the influence of uncertainty on the chosen principal components. We provide multiple examples of our technique using real-world datasets and show how to propagate multivariate normal distributions through PCA in closed-form. Furthermore, we discuss extensions and limitations of our approach.