Data Science Maturity Model (DSMM)
Many organizations have been underwhelmed by the return on their investment in data science. This is due to a narrow focus on tools, rather than a broader consideration of how data science teams work and how they fit within the larger organization. To help data science practitioners and leaders identify their existing gaps and direct future investment, Domino has developed a framework called the Data Science Maturity Model (DSMM). The DSMM assesses how reliably and sustainably a data science team can deliver value for their organization. The model consists of four levels of maturity and is split along five dimensions that apply to all analytical organizations. By design, the model is not specific to any given industry — it applies as much to an insurance company as it does to a manufacturer. …

Zolotarev Distance
In this paper the concept of a metric in the space of random variables defined on a probability space is introduced. The principle of three stages in the study of approximation problems is formulated, in particular problems of approximating distributions. Various facts connected with the use of metrics in these three stages are presented and proved. In the second part of the paper a series of results is introduced which are related to stability problems in characterizing distributions and to problems of estimating the remainder terms in limiting approximations of distributions of sums of independent random variables.
Rate of convergence to alpha stable law using Zolotarev distance : technical report …

Degree Penalty
Network embedding aims to learn the low-dimensional representations of vertexes in a network, while structure and inherent properties of the network is preserved. Existing network embedding works primarily focus on preserving the microscopic structure, such as the first- and second-order proximity of vertexes, while the macroscopic scale-free property is largely ignored. Scale-free property depicts the fact that vertex degrees follow a heavy-tailed distribution (i.e., only a few vertexes have high degrees) and is a critical property of real-world networks, such as social networks. In this paper, we study the problem of learning representations for scale-free networks. We first theoretically analyze the difficulty of embedding and reconstructing a scale-free network in the Euclidean space, by converting our problem to the sphere packing problem. Then, we propose the ‘degree penalty’ principle for designing scale-free property preserving network embedding algorithm: punishing the proximity between high-degree vertexes. We introduce two implementations of our principle by utilizing the spectral techniques and a skip-gram model respectively. Extensive experiments on six datasets show that our algorithms are able to not only reconstruct heavy-tailed distributed degree distribution, but also outperform state-of-the-art embedding models in various network mining tasks, such as vertex classification and link prediction. …