Long-tail learning is the problem of learning under skewed label distributions, which pose a challenge for standard learners. Several recent approaches for the problem have proposed enforcing a suitable margin in logit space. Such techniques are intuitive analogues of the guiding principle behind SVMs, and are equally applicable to linear models and neural models. However, when applied to neural models, such techniques do not explicitly control the geometry of the learned embeddings. This can be potentially sub-optimal, since embeddings for tail classes may be diffuse, resulting in poor generalization for these classes. We present Embedding and Logit Margins (ELM), a unified approach to enforce margins in logit space, and regularize the distribution of embeddings. This connects losses for long-tail learning to proposals in the literature on metric embedding, and contrastive learning. We theoretically show that minimising the proposed ELM objective helps reduce the generalisation gap. The ELM method is shown to perform well empirically, and results in tighter tail class embeddings.

Identification of the current and expected future pollution sources to rivers is crucial for sound environmental management. For this purpose numerous approaches were proposed that can be clustered under physical based models, stable isotope analysis and mixing methods, mass balance methods, time series analysis, land cover analysis, and spatial statistics. Another extremely common method is Principal Component Analysis, as well as its modifications, such as Absolute Principal Component Score. they have been applied to the source identification problems for nitrogen entry to rivers. This manuscript is checking whether PCA can really be a powerful method to uncover nitrogen pollution sources considering its theoretical background and assumptions. Moreover, slightly similar techniques, Independent Component Analysis and Factor Analysis will also be considered.

In this paper we extend the work of Owen (2007) by deriving a second order expansion for the slope parameter in logistic regression, when the size of the majority class is unbounded and the minority class is finite. More precisely, we demonstrate that the second order term converges to a normal distribution and explicitly compute its variance, which surprisingly once again depends only on the mean of the minority class points and not their arrangement under mild regularity assumptions. In the case that the majority class is normally distributed, we illustrate that the variance of the the limiting slope depends exponentially on the z-score of the average of the minority class's points with respect to the majority class's distribution. We confirm our results by Monte Carlo simulations.

Hypernymy plays a fundamental role in many AI tasks like taxonomy learning, ontology learning, etc. This has motivated the development of many automatic identification methods for extracting this relation, most of which rely on word distribution. We present a novel model HyperBox to learn box embeddings for hypernym discovery. Given an input term, HyperBox retrieves its suitable hypernym from a target corpus. For this task, we use the dataset published for SemEval 2018 Shared Task on Hypernym Discovery. We compare the performance of our model on two specific domains of knowledge: medical and music. Experimentally, we show that our model outperforms existing methods on the majority of the evaluation metrics. Moreover, our model generalize well over unseen hypernymy pairs using only a small set of training data.

Differentially private (DP) synthetic data generation is a practical method for improving access to data as a means to encourage productive partnerships. One issue inherent to DP is that the "privacy budget" is generally "spent" evenly across features in the data set. This leads to good statistical parity with the real data, but can undervalue the conditional probabilities and marginals that are critical for predictive quality of synthetic data. Further, loss of predictive quality may be non-uniform across the data set, with subsets that correspond to minority groups potentially suffering a higher loss. In this paper, we develop ensemble methods that distribute the privacy budget "wisely" to maximize predictive accuracy of models trained on DP data, and "fairly" to bound potential disparities in accuracy across groups and reduce inequality. Our methods are based on the insights that feature importance can inform how privacy budget is allocated, and, further, that per-group feature importance and fairness-related performance objectives can be incorporated in the allocation. These insights make our methods tunable to social contexts, allowing data owners to produce balanced synthetic data for predictive analysis.

An important step in developing machine learning models is to evaluate the performance. Depending on the type of machine learning problem that you are dealing with, there is generally a choice of metrics to choose from to perform this step. However, simply looking at one or two numbers in isolation cannot always enable us to make the right choice for model selection. For example, a single error metric doesn't give us any information about the distribution of the errors. It does not answer questions like is the model wrong in a big way a small number of times, or is it producing lots of smaller errors?

If you are on the quest for a (Supervised) Deep Learning algorithm for semantic segmentation -- keywords alert -- you certainly have found yourself searching for some high-quality labels a high quantity of data points. In our 3D data world, the unlabelled nature of the 3D point clouds makes it particularly challenging to answer both criteria: without any good training set, it is hard to "train" any predictive model. Should we explore python tricks and add them to our quiver to quickly produce awesome 3D labeled point cloud datasets? Let us dive right in! Why unsupervised segmentation & clustering is the "bulk of AI"? Deep Learning (DL) through supervised systems is extremely useful. DL architectures have profoundly changed the technological landscape in the last years.

Machine learning is increasingly shaping future of work and jobs. With an average salary of $120,000 (Glassdoor and Indeed), Machine Learning will help you to get one of the top-paying jobs. Machine Learning, provides computers the ability to automatically learn and improve from experience. Today, data scientists are generally divided among two languages, some prefer R, some prefer Python. Learning Machine Learning is a definite way to advance your career and will open doors to new Job opportunities.

Large-scale machine learning and data mining applications require computer systems to perform massive matrix-vector and matrix-matrix multiplication operations that need to be parallelized across multiple nodes. The presence of straggling nodes--computing nodes that unpredictably slow down or fail--is a major bottleneck in such distributed computations. Ideal load balancing strategies that dynamically allocate more tasks to faster nodes require knowledge or monitoring of node speeds as well as the ability to quickly move data. Recently proposed fixed-rate erasure coding strategies can handle unpredictable node slowdown, but they ignore partial work done by straggling nodes, thus resulting in a lot of redundant computation. We propose a rateless fountain coding strategy that achieves the best of both worlds--we prove that its latency is asymptotically equal to ideal load balancing, and it performs asymptotically zero redundant computations. Our idea is to create linear combinations of the m rows of the matrix and assign these encoded rows to different worker nodes. The original matrix-vector product can be decoded as soon as slightly more than m row-vector products are collectively finished by the nodes. Evaluation on parallel and distributed computing yields as much as three times speedup over uncoded schemes. Matrix-vector multiplications form the core of a plethora of scientific computing and machine learning applications that include solving partial differential equations, forward and back propagation in neural networks, computing the PageRank of graphs, etcetera. In the age of Big Data, most of these applications involve multiplying extremely large matrices and vectors and the computations cannot be performed efficiently on a single machine. This has motivated the development of several algorithms that seek to speed up matrix-vector multiplication by distributing the computation across multiple computing nodes.

We propose a novel federated learning paradigm to model data variability among heterogeneous clients in multi-centric studies. Our method is expressed through a hierarchical Bayesian latent variable model, where client-specific parameters are assumed to be realization from a global distribution at the master level, which is in turn estimated to account for data bias and variability across clients. We show that our framework can be effectively optimized through expectation maximization (EM) over latent master's distribution and clients' parameters. We also introduce formal differential privacy (DP) guarantees compatibly with our EM optimization scheme. We tested our method on the analysis of multi-modal medical imaging data and clinical scores from distributed clinical datasets of patients affected by Alzheimer's disease. We demonstrate that our method is robust when data is distributed either in iid and non-iid manners, even when local parameters perturbation is included to provide DP guarantees. Moreover, the variability of data, views and centers can be quantified in an interpretable manner, while guaranteeing high-quality data reconstruction as compared to state-of-the-art autoencoding models and federated learning schemes.