A surprising statistical phenomenon has emerged in modern machine learning: highly complex models can interpolate training data while still generalizing well to test data, even in the presence of label noise. This is rather striking as it the goes against the grain of the classical statistical wisdom which dictates that predictors that generalize well should trade off between the fit to the training data and the some measure of the complexity or smoothness of the predictor. Many estimators like neural networks, kernel estimators, nearest neighbour estimators, and even linear models have been shown to demonstrate this phenomenon (see, Zhang et al. 2017; Belkin et al. 2019, among others). This phenomenon has recently inspired intense theoretical research. One line of work (Soudry et al. 2018; Ji and Telgarsky 2019; Gunasekar et al. 2017; Nacson, Srebro, and Soudry 2019; Gunasekar et al. 2018a; Gunasekar et al. 2018b) formalized the argument (Neyshabur, Tomioka, and Srebro 2014; Neyshabur 2017) that, even when there is no explicit regularization that is used in training these rich models, there is nevertheless implicit regularization encoded in the choice of the optimization method used. For example, in the setting of linear classification, (Soudry et al. 2018; Ji and Telgarsky 2019; Nacson, Srebro, and Soudry 2019) show that learning a linear classifier using gradient descent on the unregularized logistic or exponential loss asymptotically leads the solution to converge to the maximum l

I was looking for a list of Machine Learning datasets for comparing Logistic Regression model but I couldn't find it easily. I spent some time curating it based on my need. This post is collection of such datasets which you can download for your use. The data set contains 3 classes of 50 instances each, where each class refers to a type of iris plant. One class is linearly separable from the other 2; the latter are NOT linearly separable from each other.

Coronavirus disease 2019 (COVID-19) has affected millions of people globally, yet how the human immune system responds to and influences COVID-19 severity remains unclear. Mathew et al. present a comprehensive atlas of immune modulation associated with COVID-19. They performed high-dimensional flow cytometry of hospitalized COVID-19 patients and found three prominent and distinct immunotypes that are related to disease severity and clinical parameters. Arunachalam et al. report a systems biology approach to assess the immune system of COVID-19 patients with mild-to-severe disease. These studies provide a compendium of immune cell information and roadmaps for potential therapeutic interventions. Science , this issue p. [eabc8511][1], p. [1210][2] ### INTRODUCTION Many patients with coronavirus disease 2019 (COVID-19), caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, present with severe respiratory disease requiring hospitalization and mechanical ventilation. Although most patients recover, disease is complex and case fatality can be as high as 10%. How human immune responses control or exacerbate COVID-19 is currently poorly understood, and defining the nature of immune responses during acute COVID-19 could help identify therapeutics and effective vaccines. ### RATIONALE Immune dysregulation during SARS-CoV-2 infection has been implicated in pathogenesis, but currently available data remain limited. We used high-dimensional cytometry to analyze COVID-19 patients and compare them with recovered and healthy individuals and performed integrated analysis of ~200 immune features. These data were combined with ~50 clinical features to understand how the immunology of SARS-CoV-2 infection may be related to clinical patterns, disease severity, and progression. ### RESULTS Analysis of 125 hospitalized COVID-19 patients revealed that although CD4 and CD8 T cells were activated in some patients, T cell responses were limited in others. In many patients, CD4 and CD8 T cell proliferation (measured by KI67 increase) and activation (detected by CD38 and HLA-DR coexpression) were consistent with antiviral responses observed in other infections. Plasmablast (PB) responses were present in many patients, reaching >30% of total B cells, and most patients made SARS-CoV-2–specific antibodies. However, ~20% of patients had little T cell activation or PB response compared with controls. In some patients, responses declined over time, resembling typical kinetics of antiviral responses; in others, however, robust T cell and PB responses remained stable or increased over time. These temporal patterns were associated with specific clinical features. With an unbiased uniform manifold approximation and projection (UMAP) approach, we distilled ~200 immune parameters into two major immune response components and a third pattern lacking robust adaptive immune responses, thus revealing immunotypes of COVID-19: (i) Immunotype 1 was associated with disease severity and showed robust activated CD4 T cells, a paucity of circulating follicular helper cells, activated CD8 “EMRAs,” hyperactivated or exhausted CD8 T cells, and PBs. (ii) Immunotype 2 was characterized by less CD4 T cell activation, Tbet+ effector CD4 and CD8 T cells, and proliferating memory B cells and was not associated with disease severity. (iii) Immunotype 3, which negatively correlated with disease severity and lacked obvious activated T and B cell responses, was also identified. Mortality occurred for patients with all three immunotypes, illustrating a complex relationship between immune response and COVID-19. ### CONCLUSION Three immunotypes revealing different patterns of lymphocyte responses were identified in hospitalized COVID-19 patients. These three major patterns may each represent a different suboptimal response associated with hospitalization and disease. Our findings may have implications for treatments focused on activating versus inhibiting the immune response. ![Figure][3] High-dimensional immune response analysis of COVID-19 patients identifies three immunotypes. Peripheral blood mononuclear cell immune profiling and clinical data were collected from 60 healthy donors (HDs), 36 recovered donors (RDs), and 125 hospitalized COVID-19 patients. High-dimensional flow cytometry and longitudinal analysis highlighted stability and fluctuations in the response. UMAP visualization distilled ~200 immune features into two dimensions and identified three immunotypes associated with clinical outcomes. cTfh, circulating T follicular helper cells; EMRA, a subset of effector memory T cells reexpressing CD45RA; d0, day 0. Coronavirus disease 2019 (COVID-19) is currently a global pandemic, but human immune responses to the virus remain poorly understood. We used high-dimensional cytometry to analyze 125 COVID-19 patients and compare them with recovered and healthy individuals. Integrated analysis of ~200 immune and ~50 clinical features revealed activation of T cell and B cell subsets in a proportion of patients. A subgroup of patients had T cell activation characteristic of acute viral infection and plasmablast responses reaching >30% of circulating B cells. However, another subgroup had lymphocyte activation comparable with that in uninfected individuals. Stable versus dynamic immunological signatures were identified and linked to trajectories of disease severity change. Our analyses identified three immunotypes associated with poor clinical trajectories versus improving health. These immunotypes may have implications for the design of therapeutics and vaccines for COVID-19. [1]: /lookup/doi/10.1126/science.abc8511 [2]: /lookup/doi/10.1126/science.abc6261 [3]: pending:yes

In the current tech landscape, developers are expected to have a number of different skills. And many of them do. There are also a lot of different career paths available to developers that use many of their current skills with a slight twist. Database administrators, developer advocates, and machine learning engineers all have one thing in common with all developers: they all know how to code. It doesn't matter which languages are being used, they all understand the core concepts behind writing good code. That's one of the reasons many software developers consider becoming machine learning engineers. With all of the tools and packages available, you don't need to have a deep mathematical background to get accurate results. If you are willing to learn how to use some libraries and get a high-level understanding of the underlying math, you can become a machine learning engineer.

Calibration of large-scale differential equation models to observational or experimental data is a widespread challenge throughout applied sciences and engineering. A crucial bottleneck in state-of-the art calibration methods is the calculation of local sensitivities, i.e. derivatives of the loss function with respect to the estimated parameters, which often necessitates several numerical solves of the underlying system of partial or ordinary differential equations. In this paper we present a new probabilistic approach to computing local sensitivities. The proposed method has several advantages over classical methods. Firstly, it operates within a constrained computational budget and provides a probabilistic quantification of uncertainty incurred in the sensitivities from this constraint. Secondly, information from previous sensitivity estimates can be recycled in subsequent computations, reducing the overall computational effort for iterative gradient-based calibration methods. The methodology presented is applied to two challenging test problems and compared against classical methods.

Two main routes of learning methods exist at present including neuroscience-inspired methods and machine learning methods. Both have own advantages, but neither currently can solve all learning problems well. Integrating them into one network may provide better learning abilities for general tasks. On the other hand, spiking neural network embodies "computation" in spatiotemporal domain with unique features of rich coding scheme and threshold switching, which is very suitable for low power and high parallel neuromorphic computing. Here, we report a spike-based general learning model that integrates two learning routes by introducing a brain-inspired meta-local module and a two-phase parametric modelling. The hybrid model can meta-learn general local plasticity, and receive top-down supervision information for multi-scale learning. We demonstrate that this hybrid model facilitates learning of many general tasks, including fault-tolerance learning, few-shot learning and multiple-task learning. Furthermore, the implementation of the hybrid model on the Tianjic neuromorphic platform proves that it can fully utilize the advantages of neuromorphic hardware architecture and promote energy-efficient on-chip applications.

In recent years, chat-bot has become a new type of intelligent terminal to guide users to consume services. However, it is criticized most that the services it provides are not what users expect or most expect. This defect mostly dues to two problems, one is that the incompleteness and uncertainty of user's requirement expression caused by the information asymmetry, the other is that the diversity of service resources leads to the difficulty of service selection. Conversational bot is a typical mesh device, so the guided multi-rounds Q$\&$A is the most effective way to elicit user requirements. Obviously, complex Q$\&$A with too many rounds is boring and always leads to bad user experience. Therefore, we aim to obtain user requirements as accurately as possible in as few rounds as possible. To achieve this, a user intention recognition method based on Knowledge Graph (KG) was developed for fuzzy requirement inference, and a requirement elicitation method based on Granular Computing was proposed for dialog policy generation. Experimental results show that these two methods can effectively reduce the number of conversation rounds, and can quickly and accurately identify the user intention.

Highly regulated domains such as finance have long favoured the use of machine learning algorithms that are scalable, transparent, robust and yield better performance. One of the most prominent examples of such an algorithm is XGBoost. Meanwhile, there is also a growing interest in building fair and unbiased models in these regulated domains and numerous bias-mitigation algorithms have been proposed to this end. However, most of these bias-mitigation methods are restricted to specific model families such as logistic regression or support vector machine models, thus leaving modelers with a difficult decision of choosing between fairness from the bias-mitigation algorithms and scalability, transparency, performance from algorithms such as XGBoost. We aim to leverage the best of both worlds by proposing a fair variant of XGBoost that enjoys all the advantages of XGBoost, while also matching the levels of fairness from the state-of-the-art bias-mitigation algorithms. Furthermore, the proposed solution requires very little in terms of changes to the original XGBoost library, thus making it easy for adoption. We provide an empirical analysis of our proposed method on standard benchmark datasets used in the fairness community.

We consider the problem of fitting variational posterior approximations using stochastic optimization methods. The performance of these approximations depends on (1) how well the variational family matches the true posterior distribution, (2) the choice of divergence, and (3) the optimization of the variational objective. We show that even in the best-case scenario when the exact posterior belongs to the assumed variational family, common stochastic optimization methods lead to poor variational approximations if the problem dimension is moderately large. We also demonstrate that these methods are not robust across diverse model types. Motivated by these findings, we develop a more robust and accurate stochastic optimization framework by viewing the underlying optimization algorithm as producing a Markov chain. Our approach is theoretically motivated and includes a diagnostic for convergence and a novel stopping rule, both of which are robust to noisy evaluations of the objective function. We show empirically that the proposed framework works well on a diverse set of models: it can automatically detect stochastic optimization failure or inaccurate variational approximation.

Digital technology has made possible unimaginable applications come true. It seems exciting to have a handful of tools for easy editing and manipulation, but it raises alarming concerns that can propagate as speech clones, duplicates, or maybe deep fakes. Validating the authenticity of a speech is one of the primary problems of digital audio forensics. We propose an approach to distinguish human speech from AI synthesized speech exploiting the Bi-spectral and Cepstral analysis. Higher-order statistics have less correlation for human speech in comparison to a synthesized speech. Also, Cepstral analysis revealed a durable power component in human speech that is missing for a synthesized speech. We integrate both these analyses and propose a machine learning model to detect AI synthesized speech.