In this paper, we focus on the task of learning influential parameters under unsteady and dynamic environments. Such unsteady and dynamic environments often occur in the ramp-up phase of manufacturing. We propose a novel regularization-based framework, called Distributed Dynamic Elastic Nets (DDEN), for this problem and formulate it as a convex optimization objective. Our approach solves the optimization problem using a distributed framework. Consequently it is highly scalable and can easily be applied to very large datasets. We implement a L-BFGS based solver in the Apache Spark framework. For validating our algorithm, we consider the issue of scrap reduction at an assembly line during the ramp-up phase of a manufacturing plant. By considering the logistic regression as a sample model, we evaluate the performance of our approach although extensions of the proposed regularizer to other classification and regression techniques is straightforward. Through experiments on data collected at a functioning manufacturing plant, we show that the proposed method not only reduces model variance but also helps preserve the relative importance of features in dynamic conditions compared to standard approaches. The experiments further show that the classification performance of DDEN if often better than logistic regression with standard elastic nets for datasets from dynamic and unsteady environments. We are collaborating with manufacturing units to use this process for improving production yields during the ramp-up phase. This work serves as a demonstration of how data mining can be used to solve problem in manufacturing.

We study the interplay between a dynamic process and the structure of the network on which it is defined. Specifically, we examine the impact of this interaction on the quality-measure of network clusters and node centrality. This enables us to effectively identify network communities and important nodes participating in the dynamics. As the first step towards this objective, we introduce an umbrella framework for defining and characterizing an ensemble of dynamic processes on a network. This framework generalizes the traditional Laplacian framework to continuous-time biased random walks and also allows us to model some epidemic processes over a network. For each dynamic process in our framework, we can define a function that measures the quality of every subset of nodes as a potential cluster (or community) with respect to this process on a given network. This subset-quality function generalizes the traditional conductance measure for graph partitioning. We partially justify our choice of the quality function by showing that the classic Cheeger's inequality, which relates the conductance of the best cluster in a network with a spectral quantity of its Laplacian matrix, can be extended from the Laplacian-conductance setting to this more general setting.

We present Multi-root Automatic Incremental Routing (MAIR), an efficient routing approach for mobile ad hoc networks (MANET). MAIR has a low routing stretch (ratio of selected path to shortest path length) and provides multiple paths to each destination. Every node is assigned multiple prefix labels with respect to multiple roots in the network. The roots are distributed in the network such that the paths calculated from each of the root labels are as disjoint as possible from each other. The labels of a node are stored distributively in hash tables at “anchor nodes” across the network. Data packets are routed using the distributed hash table (DHT) lookup and longest prefix match with neighbor labels. This eliminates the need to maintain large routing tables in the nodes, which substantially reduces the routing state at each node. A region of interest (ROI) is formed around each active source-destination pair using the node labels. The nodes in the ROI maintain the most recent mapping of the node identifier of a destination to its labels. This reduces the route establishment delay for the nodes inside ROI as the need for DHT lookup is reduced.

We present Multi-label Automatic Routing (MAR), the first compact routing protocol that attains a low path stretch (ratio of selected path length to the optimal path length) while maintaining a low routing state for mobile networks. MAR is resilient to node movements in the network. In MAR, nodes assign themselves labels based on their location in the network through a distributed algorithm. Distributed Hash Tables (DHTs) for the node to label mappings are established in some anchor nodes. Once the labels are established, the routing is automatic based on the positional labels of the nodes and DHT lookups. This eliminates flooding completely. Unlike traditional routing protocols MAR does not need destinations-based routing tables. Hence, MAR has a small routing state. With the use of multiple labels per node, the average path length is close to the shortest path and there are multiple paths between source and destination nodes. In Qualnet simulations MAR shows a path stretch close to or better than traditional table-driven and on-demand protocols like OLSR and AODV. Simulation results also show shorter end-to-end delays due to the automatic routing. The delivery ratio of MAR is comparable to these traditional protocols but with a significantly lower network overhead.

We present the Ordered Proactive Enclave-based Routing for Ad-hoc networks (OPERA) protocol for unicast routing in mobile ad hoc networks (MANETs). OPERA uses source and destination labels to define elliptical interest-driven enclaves to reduce control overhead. A topological sort of destination labels is used for loop freedom. The use of label spacing allows effective local repairs. Simulation results comparing OPERA with traditional routing schemes such as AODV and OLSR show that it has a better delivery ratio and smaller delay in realistic load situations. The network load with OPERA is much smaller than with the traditional on-demand and proactive routing schemes.

How does information flow in online social networks? How does the structure and size of the information cascade evolve in time? How can we efficiently mine the information contained in cascade dynamics? We approach these questions empirically and present an efficient and scalable mathematical framework for quantitative analysis of cascades on networks. We define a cascade generating function that captures the details of the microscopic dynamics of the cascades. We show that this function can also be used to compute the macroscopic properties of cascades, such as their size, spread, diameter, number of paths, and average path length. We present an algorithm to efficiently compute cascade generating function and demonstrate that while significantly compressing information within a cascade, it nevertheless allows us to accurately reconstruct its structure. We use this framework to study information dynamics on the social network of Digg. Digg allows users to post and vote on stories, and easily see the stories that friends have voted on. As a story spreads on Digg through voting, it generates cascades. We extract cascades of more than 3,500 Digg stories and calculate their macroscopic and microscopic properties. We identify several trends in cascade dynamics: spreading via chaining, branching and community. We discuss how these affect the spread of the story through the Digg social network. Our computational framework is general and offers a practical solution to quantitative analysis of the microscopic structure of even very large cascades.

We present the design and performance of a new approach to packet switching for MANETs, which we call context aware protocol engines (CAPE). With CAPE, nodes disseminate information in the network by means of context-aware packet switching that enables the statistical multiplexing of bandwidth, processing and storage resources using integrated signaling covering channel access, routing and other functions, to share and store the context within which information is disseminated. Data packet headers consist of simple pointers to their context, and elections and opportunistic reservations integrated with routing are used to attain high throughput and low channel-access delay.