Showing papers by "AT&T Labs published in 2020"

ViVo: visibility-aware mobile volumetric video streaming

Abstract: In this paper, we perform a first comprehensive study of mobile volumetric video streaming. Volumetric videos are truly 3D, allowing six degrees of freedom (6DoF) movement for their viewers during playback. Such flexibility enables numerous applications in entertainment, healthcare, education, etc. However, volumetric video streaming is extremely bandwidth-intensive. We conduct a detailed investigation of each of the following aspects for point cloud streaming (a popular volumetric data format): encoding, decoding, segmentation, viewport movement patterns, and viewport prediction. Motivated by the observations from the above study, we propose ViVo, which is to the best of our knowledge the first practical mobile volumetric video streaming system with three visibility-aware optimizations. ViVo judiciously determines the video content to fetch based on how, what and where a viewer perceives for reducing bandwidth consumption of volumetric video streaming. Our evaluations over real wireless networks (including commercial 5G), mobile devices and users indicate that ViVo can save on average 40% of data usage (up to 80%) with virtually no drop in visual quality.

On Leveraging Machine and Deep Learning for Throughput Prediction in Cellular Networks: Design, Performance, and Challenges

Abstract: The highly dynamic wireless communication environment poses a challenge for many applications (e.g., adaptive multimedia streaming services). Providing accurate TP can significantly improve performance of these applications. The scheduling algorithms in cellular networks consider various PHY metrics, (e.g., CQI) and throughput history when assigning resources for each user. This article explains how AI can be leveraged for accurate TP in cellular networks using PHY and application layer metrics. We present key architectural components and implementation options, illustrating their advantages and limitations. We also highlight key design choices and investigate their impact on prediction accuracy using real data. We believe this is the first study that examines the impact of integrating network-level data and applying a deep learning technique (on PHY and application data) for TP in cellular systems. Using video streaming as a use case, we illustrate how accurate TP improves the end user's QoE. Furthermore, we identify open questions and research challenges in the area of AI-driven TP. Finally, we report on lessons learned and provide conclusions that we believe will be useful to network practitioners seeking to apply AI.

Cellular Network Traffic Prediction Incorporating Handover: A Graph Convolutional Approach

Abstract: Cellular traffic prediction enables operators to adapt to traffic demand in real-time for improving network resource utilization and user experience. To predict cellular traffic, previous studies either applied Recurrent Neural Networks (RNN) at individual base stations or adapted Convolutional Neural Networks (CNN) to work at grid-cells in a geographically defined grid. These solutions do not consider explicitly the effect of handover on the spatial characteristics of the traffic, which may lead to lower prediction accuracy. Furthermore, RNN solutions are slow to train, and CNN-grid solutions do not work for cells and are difficult to apply to base stations. This paper proposes a new prediction model, STGCN-HO, that uses the transition probability matrix of the handover graph to improve traffic prediction. STGCN-HO builds a stacked residual neural network structure incorporating graph convolutions and gated linear units to capture both spatial and temporal aspects of the traffic. Unlike RNN, STGCN-HO is fast to train and simultaneously predicts traffic demand for all base stations based on the information gathered from the whole graph. Unlike CNN-grid, STGCN-HO can make predictions not only for base stations, but also for cells within base stations. Experiments using data from a large cellular network operator demonstrate that our model outperforms existing solutions in terms of prediction accuracy.

Data-driven domain discovery for structured datasets

Abstract: The growing number of open datasets has created new opportunities to derive insights and address important societal problems. These data, however, often come with little or no metadata, in particular about the types of their attributes, thus greatly limiting their utility. In this paper, we address the problem of domain discovery: given a collection of tables, we aim to identify sets of terms that represent instances of a semantic concept or domain. Knowledge of attribute domains not only enables a richer set of queries over dataset collections, but it can also help in data integration. We propose a data-driven approach that leverages value co-occurrence information across a large number of dataset columns to derive robust context signatures and infer domains. We discuss the results of a detailed experimental evaluation, using real urban dataset collections, which show that our approach is robust and outperforms state-of-the-art methods in the presence of incomplete columns, heterogeneous or erroneous data, and scales to datasets with several million distinct terms.

Auto-Scaling Network Service Chains Using Machine Learning and Negotiation Game

Abstract: Network Function Virtualization (NFV) enables Network Operators (NOs) to efficiently respond to the increasing dynamicity of network services. Virtual Network Functions (VNFs) running on commercial off-the-shelf servers are easy to deploy, update, monitor, and manage. Such virtualized services are often deployed as Service Chains (SCs), which require in-sequence placement of computing and memory resources as well as routing of traffic flows. Due to the ongoing migration towards cloudification of networks, the concept of auto-scaling which originated in Cloud Computing, is now receiving attention from networks professionals too. Prior studies on auto-scaling use measured load to dynamically react to traffic changes. Moreover, they often focus on only one of the resources (e.g., compute only, or network capacity only). In this study, we consider three different resource types: compute, memory, and network bandwidth. In prior studies, NO takes auto-scaling decisions, assuming tenants are always willing to auto-scale, and Quality of Service (QoS) requirements are homogeneous. Our study proposes a negotiation-game-based auto-scaling method where tenants and NO both engage in the auto-scaling decision, based on their willingness to participate, heterogeneous QoS requirements, and financial gain (e.g., cost savings). In addition, we propose a proactive Machine Learning (ML) based prediction method to perform SC auto-scaling in dynamic traffic scenario. Numerical examples show that our proposed SC auto-scaling methods powered by ML present a win-win situation for both NO and tenants (in terms of cost savings).

Characterization of Multi-User Augmented Reality over Cellular Networks

Abstract: Augmented reality (AR) apps where multiple users interact within the same physical space are gaining in popularity (e.g., shared AR mode in Pokemon Go, virtual graffiti in Google’s Just a Line). However, multi-user AR apps running over the cellular network can experience very high end-to-end latencies (measured at 12.5 s median on a public LTE network). To characterize and understand the root causes of this problem, we perform a first-of-its-kind measurement study on both public LTE and industry LTE testbed for two popular multi-user AR applications, yielding several insights: (1) The radio access network (RAN) accounts for a significant fraction of the end-to-end latency (31.2%, or 3.9 s median), resulting in AR users experiencing high, variable delays when interacting with a common set of virtual objects in off-the-shelf AR apps; (2) AR network traffic is characterized by large intermittent spikes on a single uplink TCP connection, resulting in frequent TCP slow starts that can increase user-perceived latency; (3) Applying a common traffic management mechanism of cellular operators, QoS Class Identifiers (QCI), can help by reducing AR latency by 33% but impacts non-AR users. Based on these insights, we propose network-aware and network-agnostic AR design optimization solutions to intelligently adapt IP packet sizes and periodically provide information on uplink data availability, respectively. Our solutions help ramp up network performance, improving the end-to-end AR latency and goodput by ~40-70%.

The OpenROADM initiative [Invited]

Abstract: Historically, reconfigurable optical add/drop multiplexers (ROADMs) have been proprietary with respect to their data and control planes. There has been a growing desire among network operators to move away from such proprietary elements. Together with other network operators and industry partners, we have created the OpenROADM multi-source agreement and published interoperability specifications as well as YANG data models for OpenROADM-compliant hardware and software. This paper provides an overview of the OpenROADM initiative, its current status, and the results from interoperability tests. It also includes a description of the OpenDaylight TransportPCE project for an OpenROADM-compliant software-defined networking controller.

A Control Theoretic Approach to ABR Video Streaming: A Fresh Look at PID-Based Rate Adaptation

Abstract: Adaptive bitrate streaming (ABR) has become the de facto technique for video streaming over the Internet. Despite a flurry of techniques, achieving high quality ABR streaming over cellular networks remains a tremendous challenge. ABR streaming can be naturally modeled as a control problem. There has been some initial work on using PID, a widely used feedback control technique, for ABR streaming. Existing studies, however, either use PID control directly without fully considering the special requirements of ABR streaming, leading to suboptimal results, or conclude that PID is not a suitable approach. In this paper, we take a fresh look at PID-based control for ABR streaming. We design a framework called PIA (PID-control based ABR streaming) that strategically leverages PID control concepts and incorporates several novel strategies to account for the various requirements of ABR streaming. We evaluate PIA using simulation based on real LTE network traces, as well as using real DASH implementation. The results demonstrate that PIA outperforms state-of-the-art schemes in providing high average bitrate with significantly lower bitrate changes (reduction up to 40 percent) and stalls (reduction up to 85 percent), while incurring very small runtime overhead. We further design PIA-E (PIA Enhanced), which improves the performance of PIA in the important initial playback phase.

Scaling in income inequalities and its dynamical origin

Abstract: We provide an analytically treatable model that describes in a unified manner income distribution for all income categories. The approach is based on a master equation with growth and reset terms. The model assumptions on the growth and reset rates are tested on an exhaustive database with incomes on individual level spanning a nine year period in the Cluj county (Romania). In agreement with our theoretical predictions we find that income distributions computed for several years collapse on a master-curve when a properly normalised income is considered. The Beta Prime distribution is appropriate to fit the collapsed data and it is shown that distributions derived for other countries are following similar trends with different fit parameters. The non-universal feature of the fit parameters suggests that for a more realistic modelling the model parameters have to be linked with specific socio-economic regulations.

Double-Sequence Frequency Synchronization for Wideband Millimeter-Wave Systems With Few-Bit ADCs

Abstract: In this paper, we propose and evaluate a novel double-sequence low-resolution frequency synchronization method in millimeter-wave (mmWave) systems. In our system model, the base station uses analog beams to send the synchronization signal with infinite-resolution digital-to-analog converters. The user equipment employs a fully digital front end to detect the synchronization signal with low-resolution analog-to-digital converters (ADCs). The key ingredient of the proposed method is the custom designed synchronization sequence pairs, from which there exists an invertible function (a ratio metric) of the carrier frequency offset (CFO) to be estimated. We use numerical examples to show that the ratio metric is robust to the quantization distortion. To implement our proposed method in practice, we propose to optimize the double-sequence design parameters such that: (i) for each individual user, the impact of the quantization distortion on the CFO estimation accuracy is minimized, and (ii) the resulting frequency range of estimation can capture as many users’ CFOs as possible. Numerical results reveal that our proposed algorithm can provide a flexible means to estimate CFO in a variety of low-resolution settings.

Mobile Volumetric Video Streaming Enhanced by Super Resolution

Abstract: Volumetric videos allow viewers to exercise 6-DoF (degrees of freedom) movement when watching them. Due to their true 3D nature, streaming volumetric videos is highly bandwidth demanding. In this work, we present to our knowledge a first volumetric video streaming system that leverages deep super resolution (SR) to boost the video quality on commodity mobile devices. We propose a series of judicious optimizations to make SR efficient on mobile devices.

Effective Keyword Search over Weighted Graphs

Abstract: Real graphs often contain edge and node weights, representing, for instance, penalty, distance or uncertainty. We study the problem of keyword search over weighted node-labeled graphs, in which a query consists of a set of keywords and an answer is a subgraph whose nodes contain the keywords. We evaluate answers using three ranking strategies: optimizing edge weights, optimizing node weights, and a bi-objective combination of both node and edge weights. We prove that optimizing node weights and the bi-objective function are NP-hard. We propose an algorithm that optimizes edge weights and has an approximation ratio of two for the unique node enumeration paradigm. To optimize node weights and the bi-objective function, we propose transformations that distribute node weights onto the edges. We then prove that our transformations allow our algorithm to also optimize node weights and the bi-objective function with the same approximation ratio of two. Notably, the proposed transformations are compatible with existing algorithms that only optimize edge weights. We empirically show that in many natural examples, incorporating node weights (both keyword holders and middle nodes) produces more relevant answers than ranking methods based only on edge weights. Extensive experiments over real-life datasets verify the effectiveness and efficiency of our solution.

Energy considerations for ABR video streaming to smartphones: measurements, models and insights

Abstract: Adaptive Bitrate (ABR) streaming is widely used in commercial video services. In this paper, we profile energy consumption of ABR streaming on mobile devices. This profiling is important, since the insights can help developing more energy-efficient ABR streaming pipelines and techniques. We first develop component power models that provide online estimation of the power draw for each component involved in ABR streaming. Using these models, we then quantify the power breakdown in ABR streaming for both regular videos and the emerging 360° panoramic videos. Our measurements validate the accuracy of the power models and provide a number of insights. We discuss use cases of the developed power models, and explore two energy reduction strategies for ABR streaming. Evaluation demonstrates that these simple strategies can provide up to 30% energy savings, with little degradation in viewing quality.

Slimmer: Accelerating 3D Semantic Segmentation for Mobile Augmented Reality

Abstract: Three-Dimensional (3D) semantic segmentation is an essential building block for interactive Augmented Reality (AR). However, existing Deep Neural Network (DNN) models for segmenting 3D objects are not only computation-intensive but also memory heavy, hindering their deployment on resourceconstrained mobile devices. We present the design, implementation and evaluation of Slimmer, a generic and model-independent framework for accelerating 3D semantic segmentation and facilitating its real-time applications on mobile devices. In contrast to the current practice that directly feeds a point cloud to DNN models, Slimmer is motivated by our observation that these models remain high accuracy even if we remove a fraction of points from the input, which can significantly reduce the inference time and memory usage of these models. Our design of Slimmer faces two key challenges. First, the simplification method of point clouds should be lightweight. Otherwise, the reduced inference time may be canceled out by the incurred overhead of input-data simplification. Second, Slimmer still needs to accurately segment the removed points from the input to create a complete segmentation of the original input, again, using a lightweight method. Our extensive performance evaluation demonstrates that, by addressing these two challenges, Slimmer can dramatically reduce the resource utilization of a representative DNN model for 3D semantic segmentation. For example, if we can tolerate 1% accuracy loss, the reduction could be $\sim$20% for inference time and$\sim$9% for memory usage. The reduction increases to around $\sim$27% for inference time and$\sim$15% for memory usage when we can tolerate 2% accuracy loss.

Natural Key Discovery in Wikipedia Tables

Abstract: Wikipedia is the largest encyclopedia to date. Scattered among its articles, there is an enormous number of tables that contain structured, relational information. In contrast to database tables, these webtables lack metadata, making it difficult to automatically interpret the knowledge they harbor. The natural key is a particularly important piece of metadata, which acts as a primary key and consists of attributes inherent to an entity. Determining natural keys is crucial for many tasks, such as information integration, table augmentation, or tracking changes to entities over time. To address this challenge, we formally define the notion of natural keys and propose a supervised learning approach to automatically detect natural keys in Wikipedia tables using carefully engineered features. Our solution includes novel features that extract information from time (a table’s version history) and space (other similar tables). On a curated dataset of 1,000 Wikipedia table histories, our model achieves 80% F-measure, which is at least 20% more than all related approaches. We use our model to discover natural keys in the entire corpus of Wikipedia tables and provide the dataset to the community to facilitate future research.

A Study of Network-Side 5G User Localization Using Angle-Based Fingerprints

Abstract: This paper explores network-side cellular user localization using fingerprints created from the angle measurements enabled by 5G. Our key idea is a binning-based fingerprinting technique that leverages multipath propagation to create fingerprint vectors based on angles of arrival of signals along multiple paths at each user. In network simulations that recreate urban environments with 3D building geometry and base station locations for a major city, our binning-based fingerprinting for 5G achieves significantly lower localization errors with a single base station than signal strength-based fingerprinting for LTE.

SessionStore: A Session-Aware Datastore for the Edge

Abstract: It is common for storage systems designed to run on edge datacenters to avoid the high latencies associated with geo-distribution by relying on eventually consistent models to replicate data. Eventual consistency works well for many edge applications because as long as the client interacts with the same replica, the storage system can provide session consistency, a stronger consistency model that has two additional important properties: (i) read-your-writes, where subsequent reads by a client that has updated an object will return the updated value or a newer one; and, (ii) monotonic reads, where if a client has seen a particular value for an object, subsequent reads will return the same value or a newer one. While session consistency does not guarantee that different clients will perceive updates in the same order, it nevertheless presents each individual client with an intuitive view of the world that is consistent with the client’s own actions. Unfortunately, these consistency guarantees break down when a client interacts with multiple replicas housed on different datacenters over time, either as a result of application partitioning, or client or code mobility.SessionStore is a datastore for fog/edge computing that ensures session consistency on a top of otherwise eventually consistent replicas. SessionStore enforces session consistency by grouping related data accesses into a session, and using a session-aware reconciliation algorithm to reconcile only the data that is relevant to the session when switching between replicas. This approach reduces data transfer and latency by up to 90% compared to full replica reconciliation.

Real-time Millimeter Wave Omnidirectional Channel Sounder Using Phased Array Antennas

Abstract: Characterization of the millimeter wave wireless channel is needed to facilitate fully connected vehicular communication in the future. To study the multipath-rich, rapidly varying nature of the vehicular propagation environment, fast millimeter wave channel sounders are required. We present a channel sounder design capable of covering 360 degrees in azimuth and 60 degrees in elevation with 200 individual beam directions in 6.25 ms by using four phased arrays simultaneously. The channel measurements are accompanied by high resolution positioning and video data, allowing channel sounding to be conducted while either the transmitter, or the receiver, or both are moving. Channel sounding campaigns were conducted at multiple urban locations with light traffic conditions in Austin, Texas. Preliminary results show that beam selection at the receiver can lower the effective pathloss exponent to 1.6 for line-of-sight and 2.25 for non line-of-sight.

Experience: towards automated customer issue resolution in cellular networks

Abstract: Cellular service carriers often employ reactive strategies to assist customers who experience non-outage related individual service degradation issues (e.g., service performance degradations that do not impact customers at scale and are likely caused by network provisioning issues for individual devices). Customers need to contact customer care to request assistance before these issues are resolved. This paper presents our experience with PACE (ProActive customer CarE), a novel, proactive system that monitors, troubleshoots and resolves individual service issues, without having to rely on customers to first contact customer care for assistance. PACE seeks to improve customer experience and care operation efficiency by automatically detecting individual (non-outage related) service issues, prioritizing repair actions by predicting customers who are likely to contact care to report their issues, and proactively triggering actions to resolve these issues. We develop three machine learning-based prediction models, and implement a fully automated system that integrates these prediction models and takes resolution actions for individual customers. We conduct a large-scale trace-driven evaluation using real-world data collected from a major cellular carrier in the US, and demonstrate that PACE is able to predict customers who are likely to contact care due to non-outage related individual service issues with high accuracy. We further deploy PACE into this cellular carrier network. Our field trial results show that PACE is effective in proactively resolving non-outage related individual customer service issues, improving customer experience, and reducing the need for customers to report their service issues.

SiEVE: Semantically Encoded Video Analytics on Edge and Cloud

Abstract: Recent advances in computer vision and neural networks have made it possible for more surveillance videos to be automatically searched and analyzed by algorithms rather than humans. This happened in parallel with advances in edge computing where videos are analyzed over hierarchical clusters that contain edge devices, close to the video source. However, the current video analysis pipeline has several disadvantages when dealing with such advances. For example, video encoders have been designed for a long time to please human viewers and be agnostic of the downstream analysis task (e.g., object detection). Moreover, most of the video analytics systems leverage 2-tier architecture where the encoded video is sent to either a remote cloud or a private edge server but does not efficiently leverage both of them. In response to these advances, we present SIEVE, a 3-tier video analytics system to reduce the latency and increase the throughput of analytics over video streams. In SIEVE, we present a novel technique to detect objects in compressed video streams. We refer to this technique as semantic video encoding because it allows video encoders to be aware of the semantics of the downstream task (e.g., object detection). Our results show that by leveraging semantic video encoding, we achieve close to 100% object detection accuracy with decompressing only 3.5% of the video frames which results in more than 100x speedup compared to classical approaches that decompress every video frame.

Discovering Band Order Dependencies

Abstract: We introduce band ODs to model the semantics of attributes that are monotonically related with small variations without there being an intrinsic violation of semantics. To make band ODs relevant to real-world applications, we make them less strict to hold approximately with some exceptions. Since formulating integrity constraints manually is cumbersome, we study the problem of automatic approximate band OD discovery. We devise an algorithm that determines the optimal solution in polynomial time. We perform a thorough experimental evaluation of our techniques over real-world and synthetic datasets.

Demonstration of Joint Operation across OpenROADM Metro Network, OpenFlow Packet Domain, and OpenStack Compute Domain

Abstract: Progress on the recent implementation of OpenROADM MSA functionalities is reported along with a description of the related TransportPCE SDN controller and PROnet multi-domain resource orchestrator software modules. These functionalities enable the described use cases. © 2020 The Author(s)

Toward Mobile 3D Vision

Abstract: In the past few years, the computer vision community has developed numerous novel technologies of 3D vision (e.g., 3D object detection and classification and 3D scene segmentation). In this work, we explore the opportunities brought by these innovations for enabling real-time 3D vision on mobile devices. Mobile 3D vision finds various use cases for emerging applications such as autonomous driving, drone navigation, and augmented reality (AR). The key differences between 3D vision and 2D vision mainly stem from the input data format (i.e., point clouds or 3D meshes vs. 2D images). Hence, the key challenge of 3D vision is that it is could be more computation intensive and memory hungry than 2D vision, due to the additional dimension of input data. For example, our preliminary measurement study of several state-of-the-art machine learning models for 3D vision shows that none of them can execute faster than one frame per second on smartphones. Motivated by these challenges, we present in this position paper a research agenda on offering systems support for real-time mobile 3D vision, focusing on improving its computation efficiency and memory utilization.

Random Sampling for Group-By Queries

Abstract: Random sampling has been widely used in approximate query processing on large databases, due to its potential to significantly reduce resource usage and response times, at the cost of a small approximation error. We consider random sampling for answering the ubiquitous class of group-by queries, which first group data according to one or more attributes, and then aggregate within each group after filtering through a predicate. The challenge with group-by queries is that a sampling method cannot focus on optimizing the quality of a single answer (e.g. the mean of selected data), but must simultaneously optimize the quality of a set of answers (one per group).We present CVOPT, a query- and data-driven sampling framework for a set of group-by queries. To evaluate the quality of a sample, CVOPT defines a metric based on the norm (e.g. l 2 or l ∞ ) of the coefficients of variation (CVs) of different answers, and constructs a stratified sample that provably optimizes the metric. CVOPT can handle group-by queries on data where groups have vastly different statistical characteristics, such as frequencies, means, or variances. CVOPT jointly optimizes for multiple aggregations and multiple group-by clauses, and provides a way to prioritize specific groups or aggregates. It can be tuned to cases when partial information about a query workload is known, such as a data warehouse where queries are run periodically.Our experimental results show that CVOPT outperforms the current state-of-the-art on sample quality and estimation accuracy for group-by queries. On a set of queries on two real-world data sets, CVOPT yields relative errors that are 5× smaller than competing approaches, under the same space budget.

Summarizing Hierarchical Multidimensional Data

Abstract: Data scientists typically analyze and extract insights from large multidimensional data sets such as US census data, enterprise sales data, and so on. But before sophisticated machine learning and statistical methods are employed, it is useful to build and explore concise summaries of the data set. While a variety of summaries have been proposed over the years, the goal of creating a concise summary of multidimensional data that can provide worst-case accuracy guarantees has remained elusive. In this paper, we propose Tree Summaries, which attain this challenging goal over arbitrary hierarchical multidimensional data sets. Intuitively, a Tree Summary is a weighted "embedded tree" in the lattice that is the cross-product of the dimension hierarchies; individual data values can be efficiently estimated by looking up the weight of their unique closest ancestor in the Tree Summary. We study the problems of generating lossless as well as (given a desired worst-case accuracy guarantee a) lossy Tree Summaries. We develop a polynomial-time algorithm that constructs the optimal (i.e., most concise) Tree Summary for each of these problems; this is a surprising result given the NP-hardness of constructing a variety of other optimal summaries over multidimensional data. We complement our analytical results with an empirical evaluation of our algorithm, and demonstrate with a detailed set of experiments on real and synthetic data sets that our algorithm outperforms prior methods in terms of conciseness of summaries or accuracy of estimation.

Interactive Testing of Line-of-Sight and Fresnel Zone Clearance for Planning Microwave Backhaul Links and 5G Networks

Abstract: The growing demand for high-speed networks is increasing the use of high-frequency electromagnetic waves in wireless networks, including in microwave backhaul links and 5G. The relative higher frequency provides a high bandwidth, but it is very sensitive to obstructions and interference. Hence, when positioning a transmitter-receiver pair, the line-of-sight between them should be free of obstacles. Furthermore, the Fresnel zone around the line-of-sight should be clear of obstructions, to guarantee effective transmission. When deploying microwave backhaul links or a cellular network there is a need to select the locations of the antennas accordingly. To help network planners, we developed an interactive tool that allows users to position antennas in different locations over a 3D model of the world. Users can interactively change antenna locations and other parameters, to examine clearance of Fresnel zones. In this paper we illustrate the interactive tool and the ability to test clearance in real-time, to support interactive network planning.

Local dampening: differential privacy for non-numeric queries via local sensitivity

Abstract: Differential privacy is the state-of-the-art formal definition for data release under strong privacy guarantees. A variety of mechanisms have been proposed in the literature for releasing the noisy output of numeric queries (e.g., using the Laplace mechanism), based on the notions of global sensitivity and local sensitivity. However, although there has been some work on generic mechanisms for releasing the output of non-numeric queries using global sensitivity (e.g., the Exponential mechanism), the literature lacks generic mechanisms for releasing the output of non-numeric queries using local sensitivity to reduce the noise in the query output. In this work, we remedy this shortcoming and present the local dampening mechanism. We adapt the notion of local sensitivity for the non-numeric setting and leverage it to design a generic non-numeric mechanism. We illustrate the effectiveness of the local dampening mechanism by applying it to two diverse problems: (i) Influential node analysis. Given an influence metric, we release the top-k most central nodes while preserving the privacy of the relationship between nodes in the network; (ii) Decision tree induction. We provide a private adaptation to the ID3 algorithm to build decision trees from a given tabular dataset. Experimental results show that we could reduce the use of privacy budget by 3 to 4 orders of magnitude for Influential node analysis and increase accuracy up to 12% for Decision tree induction when compared to global sensitivity based approaches.

Large-Scale Geospatial Planning of Wireless Backhaul Links

Abstract: In telecommunication networks, microwave backhaul links are often used as wireless connections between towers. They are used in places where deploying optical fibers is impossible or too expensive. The relatively high frequency of microwaves increases their ability to transfer information at a high rate, but it also makes them susceptible to obstructions and interference. When deploying microwave links, there should be a clear line of sight between every pair of receiver and transmitter, and a buffer around the line of sight defined by the first Fresnel zone should be clear of obstacles. In this paper we discuss the geospatial aspects of microwave backhaul planning and the challenges in developing a system for large scale planning, with the following requirements: (1) the need to cover all of the USA, (2) distance of up to 80 kilometers between towers, and (3) computing batches of thousands of pairs within a few minutes.

Characterizing and Learning Representation on Customer Contact Journeys in Cellular Services

Abstract: Corporations spend billions of dollars annually caring for customers across multiple contact channels. A customer journey is the complete sequence of contacts that a given customer has with a company across multiple channels of communication. While each contact is important and contains rich information, studying customer journeys provides a better context to understand customers' behavior in order to improve customer satisfaction and loyalty, and to reduce care costs. However, journey sequences have a complex format due to the heterogeneity of user behavior: they are variable-length, multi-attribute, and exhibit a large cardinality in categories (e.g. contact reasons). The question of how to characterize and learn representations of customer journeys has not been studied in the literature. We propose to learn journey embeddings using a sequence-to-sequence framework that converts each customer journey into a fixed-length latent embedding. In order to improve the disentanglement and distributional properties of embeddings, the model is further modified by incorporating a Wasserstein autoencoder inspired regularization on the distribution of embeddings. Experiments conducted on an enterprise-scale dataset demonstrate the effectiveness of the proposed model and reveal significant improvements due to the regularization in both distinguishing journey pattern characteristics and predicting future customer engagement.

Erratum for Discovering Order Dependencies through Order Compatibility (EDBT 2019).

Abstract: A number of extensions to the classical notion of functional dependencies have been proposed to express and enforce application semantics. One of these extensions is that of order dependencies (ODs), which express rules involving order. The article entitled “Discovering Order Dependencies through Order Compatibility” by Consonni et al., published in the EDBT conference proceedings in March 2019, investigates the OD discovery problem. The authors claim to prove that their OD discovery algorithm, OCDDISCOVER, is complete, as well as being significantly more efficient in practice than the state-of-the-art. They further claim that the implementation of the existing FASTOD algorithm (ours)—we shared our code base with the authors—which they benchmark against is flawed, as OCDDISCOVER and FASTOD report different sets of ODs over the same data sets. In this rebuttal, we show that their claim of completeness is, in fact, not true. OCDDISCOVER’s pruning rules are overly aggressive, and prune parts of the search space that contain legitimate ODs. This is the reason their approach appears to be “faster” in practice. Finally, we show that Consonni et al. misinterpret our set-based canonical form for ODs, leading to an incorrect claim that our FASTOD implementation has an error.