In federated learning (FL) systems, e.g., wireless networks, the communication cost between the clients and the central server can often be a bottleneck. To reduce the communication cost, the paradigm of communication compression has become a popular strategy in the literature. In this paper, we focus on biased gradient compression techniques in non-convex FL problems. In the classical setting of distributed learning, the method of error feedback (EF) is a common technique to remedy the downsides of biased gradient compression. In this work, we study a compressed FL scheme equipped with error feedback, named Fed-EF. We further propose two variants: Fed-EF-SGD and Fed-EF-AMS, depending on the choice of the global model optimizer. We provide a generic theoretical analysis, which shows that directly applying biased compression in FL leads to a non-vanishing bias in the convergence rate. The proposed Fed-EF is able to match the convergence rate of the full-precision FL counterparts under data heterogeneity with a linear speedup. Moreover, we develop a new analysis of the EF under partial client participation, which is an important scenario in FL. We prove that under partial participation, the convergence rate of Fed-EF exhibits an extra slow-down factor due to a so-called ``stale error compensation'' effect. A numerical study is conducted to justify the intuitive impact of stale error accumulation on the norm convergence of Fed-EF under partial participation. Finally, we also demonstrate that incorporating the two-way compression in Fed-EF does not change the convergence results. In summary, our work conducts a thorough analysis of the error feedback in federated non-convex optimization. Our analysis with partial client participation also provides insights on a theoretical limitation of the error feedback mechanism, and possible directions for improvements.

Objectives: Early detection of liver fibrosis can help cure the disease or prevent disease progression. We perform a comprehensive study of machine learning-based fibrosis detection in CT images using radiomic features to develop a non-invasive approach to fibrosis detection. Methods: Two sets of radiomic features were extracted from spherical ROIs in CT images of 182 patients who underwent simultaneous liver biopsy and CT examinations, one set corresponding to biopsy locations and another distant from biopsy locations. Combinations of contrast, normalization, machine learning model, feature selection method, bin width, and kernel radius were investigated, each of which were trained and evaluated 100 times with randomized development and test cohorts. The best settings were evaluated based on their mean test AUC and the best features were determined based on their frequency among the best settings. Results: Logistic regression models with NC images normalized using Gamma correction with $\gamma = 1.5$ performed best for fibrosis detection. Boruta was the best for radiomic feature selection method. Training a model using these optimal settings and features consisting of first order energy, first order kurtosis, and first order skewness, resulted in a model that achieved mean test AUCs of 0.7549 and 0.7166 on biopsy-based and non-biopsy ROIs respectively, outperforming a baseline and best models found during the initial study. Conclusions: Logistic regression models trained on radiomic features from NC images normalized using Gamma correction with $\gamma = 1.5$ that underwent Boruta feature selection are effective for liver fibrosis detection. Energy, kurtosis, and skewness are particularly effective features for fibrosis detection.

This paper proposes a method for measuring fairness through equality of effort by applying algorithmic recourse through minimal interventions. Equality of effort is a property that can be quantified at both the individual and the group level. It answers the counterfactual question: what is the minimal cost for a protected individual or the average minimal cost for a protected group of individuals to reverse the outcome computed by an automated system? Algorithmic recourse increases the flexibility and applicability of the notion of equal effort: it overcomes its previous limitations by reconciling multiple treatment variables, introducing feasibility and plausibility constraints, and integrating the actual relative costs of interventions. We extend the existing definition of equality of effort and present an algorithm for its assessment via algorithmic recourse. We validate our approach both on synthetic data and on the German credit dataset.

As the scope of machine learning broadens, we observe a recurring theme of algorithmic monoculture: the same systems, or systems that share components (e.g. training data), are deployed by multiple decision-makers. While sharing offers clear advantages (e.g. amortizing costs), does it bear risks? We introduce and formalize one such risk, outcome homogenization: the extent to which particular individuals or groups experience negative outcomes from all decision-makers. If the same individuals or groups exclusively experience undesirable outcomes, this may institutionalize systemic exclusion and reinscribe social hierarchy. To relate algorithmic monoculture and outcome homogenization, we propose the component-sharing hypothesis: if decision-makers share components like training data or specific models, then they will produce more homogeneous outcomes. We test this hypothesis on algorithmic fairness benchmarks, demonstrating that sharing training data reliably exacerbates homogenization, with individual-level effects generally exceeding group-level effects. Further, given the dominant paradigm in AI of foundation models, i.e. models that can be adapted for myriad downstream tasks, we test whether model sharing homogenizes outcomes across tasks. We observe mixed results: we find that for both vision and language settings, the specific methods for adapting a foundation model significantly influence the degree of outcome homogenization. We conclude with philosophical analyses of and societal challenges for outcome homogenization, with an eye towards implications for deployed machine learning systems.

Plotting a learner's generalization performance against the training set size results in a so-called learning curve. This tool, providing insight in the behavior of the learner, is also practically valuable for model selection, predicting the effect of more training data, and reducing the computational complexity of training. We set out to make the (ideal) learning curve concept precise and briefly discuss the aforementioned usages of such curves. The larger part of this survey's focus, however, is on learning curves that show that more data does not necessarily leads to better generalization performance. A result that seems surprising to many researchers in the field of artificial intelligence. We point out the significance of these findings and conclude our survey with an overview and discussion of open problems in this area that warrant further theoretical and empirical investigation.

Differential privacy is the standard privacy definition for performing analyses over sensitive data. Yet, its privacy budget bounds the number of tasks an analyst can perform with reasonable accuracy, which makes it challenging to deploy in practice. This can be alleviated by private sketching, where the dataset is compressed into a single noisy sketch vector which can be shared with the analysts and used to perform arbitrarily many analyses. However, the algorithms to perform specific tasks from sketches must be developed on a case-by-case basis, which is a major impediment to their use. In this paper, we introduce the generic moment-to-moment (M$^2$M) method to perform a wide range of data exploration tasks from a single private sketch. Among other things, this method can be used to estimate empirical moments of attributes, the covariance matrix, counting queries (including histograms), and regression models. Our method treats the sketching mechanism as a black-box operation, and can thus be applied to a wide variety of sketches from the literature, widening their ranges of applications without further engineering or privacy loss, and removing some of the technical barriers to the wider adoption of sketches for data exploration under differential privacy. We validate our method with data exploration tasks on artificial and real-world data, and show that it can be used to reliably estimate statistics and train classification models from private sketches.

When several models have similar training scores, classical model selection heuristics follow Occam's razor and advise choosing the ones with least capacity. Yet, modern practice with large neural networks has often led to situations where two networks with exactly the same number of parameters score similar on the training set, but the deeper one generalizes better to unseen examples. With this in mind, it is well accepted that deep networks are superior to shallow wide ones. However, theoretically there is no difference between the two. In fact, they are both universal approximators. In this work we propose a new unsupervised measure that predicts how well a model will generalize. We call it the output sharpness, and it is based on the fact that, in reality, boundaries between concepts are generally unsharp. We test this new measure on several neural network settings, and architectures, and show how generally strong the correlation is between our metric, and test set performance. Having established this measure, we give a mathematical probabilistic argument that predicts network depth to be correlated with our proposed measure. After verifying this in real data, we are able to formulate the key argument of the work: output sharpness hampers generalization; deep networks have an in built bias against it; therefore, deep networks beat wide ones. All in all the work not only provides a helpful predictor of overfitting that can be used in practice for model selection (or even regularization), but also provides a much needed theoretical grounding for the success of modern deep neural networks.

Time series data is being used everywhere, from sales records to patients' health evolution metrics. The ability to deal with this data has become a necessity, and time series analysis and forecasting are used for the same. Every Machine Learning enthusiast would consider these as very important tools, as they deepen the understanding of the characteristics of data. Forecasting is used to predict the value of a variable in the future, based on its past occurrences. A detailed survey of the various methods that are used for forecasting has been presented in this paper. The complete process of forecasting, from preprocessing to validation has also been explained thoroughly. Various statistical and deep learning models have been considered, notably, ARIMA, Prophet and LSTMs. Hybrid versions of Machine Learning models have also been explored and elucidated. Our work can be used by anyone to develop a good understanding of the forecasting process, and to identify various state of the art models which are being used today.

In multi-label classification, each example in a dataset may be annotated as belonging to one or more classes (or none of the classes). Example applications include image (or document) tagging where each possible tag either applies to a particular image (or document) or not. With many possible classes to consider, data annotators are likely to make errors when labeling such data in practice. Here we consider algorithms for finding mislabeled examples in multi-label classification datasets. We propose an extension of the Confident Learning framework to this setting, as well as a label quality score that ranks examples with label errors much higher than those which are correctly labeled. Both approaches can utilize any trained classifier. After demonstrating that our methodology empirically outperforms other algorithms for label error detection, we apply our approach to discover many label errors in the CelebA image tagging dataset.

Maneuvering target tracking is a challenging problem for sensor systems because of the unpredictability of the targets' motions. This paper proposes a novel data-driven method for learning the dynamical motion model of a target. Non-parametric Gaussian process regression (GPR) is used to learn a target's naturally shift invariant motion (NSIM) behavior, which is translationally invariant and does not need to be constantly updated as the target moves. The learned Gaussian processes (GPs) can be applied to track targets within different surveillance regions from the surveillance region of the training data by being incorporated into the particle filter (PF) implementation. The performance of our proposed approach is evaluated over different maneuvering scenarios by being compared with commonly used interacting multiple model (IMM)-PF methods and provides around $90\%$ performance improvement for a multi-target tracking (MTT) highly maneuvering scenario.