Streaming Setting Research Articles

Neural Circuits for Dynamics-Based Segmentation of Time Series.

The brain must extract behaviorally relevant latent variables from the signals streamed by the sensory organs. Such latent variables are often encoded in the dynamics that generated the signal rather than in the specific realization of the waveform. Therefore, one problem faced by the brain is to segment time series based on underlying dynamics. We present two algorithms for performing this segmentation task that are biologically plausible, which we define as acting in a streaming setting and all learning rules being local. One algorithm is model based and can be derived from an optimization problem involving a mixture of autoregressive processes. This algorithm relies on feedback in the form of a prediction error and can also be used for forecasting future samples. In some brain regions, such as the retina, the feedback connections necessary to use the prediction error for learning are absent. For this case, we propose a second, model-free algorithm that uses a running estimate of the autocorrelation structure of the signal to perform the segmentation. We show that both algorithms do well when tasked with segmenting signals drawn from autoregressive models with piecewise-constant parameters. In particular, the segmentation accuracy is similar to that obtained from oracle-like methods in which the ground-truth parameters of the autoregressive models are known. We also test our methods on data sets generated by alternating snippets of voice recordings. We provide implementations of our algorithms at https://github.com/ttesileanu/bio-time-series.

Exploiting Smart Meter Power Consumption Measurements for Human Activity Recognition (HAR) with a Motif-Detection-Based Non-Intrusive Load Monitoring (NILM) Approach.

Numerous approaches exist for disaggregating power consumption data, referred to as non-intrusive load monitoring (NILM). Whereas NILM is primarily used for energy monitoring, we intend to disaggregate a household’s power consumption to detect human activity in the residence. Therefore, this paper presents a novel approach for NILM, which uses pattern recognition on the raw power waveform of the smart meter measurements to recognize individual household appliance actions. The presented NILM approach is capable of (near) real-time appliance action detection in a streaming setting, using edge computing. It is unique in our approach that we quantify the disaggregating uncertainty using continuous pattern correlation instead of binary device activity states. Further, we outline using the disaggregated appliance activity data for human activity recognition (HAR). To evaluate our approach, we use a dataset collected from actual households. We show that the developed NILM approach works, and the disaggregation quality depends on the pattern selection and the appliance type. In summary, we demonstrate that it is possible to detect human activity within the residence using a motif-detection-based NILM approach applied to smart meter measurements.

Prefix Imputation of Orphan Events in Event Stream Processing.

In the context of process mining, event logs consist of process instances called cases. Conformance checking is a process mining task that inspects whether a log file is conformant with an existing process model. This inspection is additionally quantifying the conformance in an explainable manner. Online conformance checking processes streaming event logs by having precise insights into the running cases and timely mitigating non-conformance, if any. State-of-the-art online conformance checking approaches bound the memory by either delimiting storage of the events per case or limiting the number of cases to a specific window width. The former technique still requires unbounded memory as the number of cases to store is unlimited, while the latter technique forgets running, not yet concluded, cases to conform to the limited window width. Consequently, the processing system may later encounter events that represent some intermediate activity as per the process model and for which the relevant case has been forgotten, to be referred to as orphan events. The naïve approach to cope with an orphan event is to either neglect its relevant case for conformance checking or treat it as an altogether new case. However, this might result in misleading process insights, for instance, overestimated non-conformance. In order to bound memory yet effectively incorporate the orphan events into processing, we propose an imputation of missing-prefix approach for such orphan events. Our approach utilizes the existing process model for imputing the missing prefix. Furthermore, we leverage the case storage management to increase the accuracy of the prefix prediction. We propose a systematic forgetting mechanism that distinguishes and forgets the cases that can be reliably regenerated as prefix upon receipt of their future orphan event. We evaluate the efficacy of our proposed approach through multiple experiments with synthetic and three real event logs while simulating a streaming setting. Our approach achieves considerably higher realistic conformance statistics than the state of the art while requiring the same storage.

Fuzzy association analysis for identifying climatic and socio-demographic factors impacting the spread of COVID-19.

Recently, the whole world witnessed the fatal outbreak of COVID-19 epidemic originating at Wuhan, Hubei province, China, during a mass gathering in a film festival. World Health Organization (WHO) has declared this COVID-19 as a pandemic due to its rapid spread across different countries within a few days. Several research works are being performed to understand the various influential factors responsible for spreading COVID. However, limited studies have been performed on how climatic and socio-demographic conditions may impact the spread of the virus. In this work, we aim to find the relationship of socio-demographic conditions, such as temperature, humidity, and population density of the regions, with the spread of COVID-19. The COVID data for different countries along with the social data are collected. For the experimental purpose, Fuzzy association rule mining is employed to infer the various relationships from the data. Moreover, to examine the seasonal effect, a streaming setting is also considered. The experimental results demonstrate various interesting insights to understand the impact of different factors on spreading COVID-19.

AFQN: approximate Qn estimation in data streams

We present afqn (Approximate Fast Qn), a novel algorithm for approximate computation of the Qn scale estimator in a streaming setting, in the sliding window model. It is well-known that computing the Qn estimator exactly may be too costly for some applications, and the problem is a fortiori exacerbated in the streaming setting, in which the time available to process incoming data stream items is short. In this paper we show how to efficiently and accurately approximate the Qn estimator. As an application, we show the use of afqn for fast detection of outliers in data streams. In particular, the outliers are detected in the sliding window model, with a simple check based on the Qn scale estimator. Extensive experimental results on synthetic and real datasets confirm the validity of our approach by showing up to three times faster updates per second. Our contributions are the following ones: (i) to the best of our knowledge, we present the first approximation algorithm for online computation of the Qn scale estimator in a streaming setting and in the sliding window model; (ii) we show how to take advantage of our UDDSketch algorithm for quantile estimation in order to quickly compute the Qn scale estimator; (iii) as an example of a possible application of the Qn scale estimator, we discuss how to detect outliers in an input data stream.

Learning from evolving data streams through ensembles of random patches

Ensemble methods represent an effective way to solve supervised learning problems. Such methods are prevalent for learning from evolving data streams. One of the main reasons for such popularity is the possibility of incorporating concept drift detection and recovery strategies in conjunction with the ensemble algorithm. On top of that, successful ensemble strategies, such as bagging and random forest, can be easily adapted to a streaming setting. In this work, we analyse a novel ensemble method designed specially to cope with evolving data streams, namely the streaming random patches (SRP) algorithm. SRP combines random subspaces and online bagging to achieve competitive predictive performance in comparison with other methods. We significantly extend previous theoretical insights and empirical results illustrating different aspects of SRP. In particular, we explain how the widely adopted incremental Hoeffding trees are not, in fact, unstable learners, unlike their batch counterparts, and how this fact significantly influences ensemble methods design and performance. We compare SRP against state-of-the-art ensemble variants for streaming data in a multitude of datasets. The results show how SRP produces a high predictive performance for both real and synthetic datasets. We also show how ensembles of random subspaces can be an efficient and accurate option to SRP and leveraging bagging as we increase the number of base learners. Besides, we analyse the diversity over time and the average tree depth, which provides insights on the differences between local subspace randomization (as in random forest) and global subspace randomization (as in random subspaces). Finally, we analyse the behaviour of SRP when using Naive Bayes as its base learner instead of Hoeffding trees.

AugSplicing: Synchronized Behavior Detection in Streaming Tensors

How can we track synchronized behavior in a stream of time-stamped tuples, such as mobile devices installing and uninstalling applications in the lockstep, to boost their ranks in the app store? We model such tuples as entries in a streaming tensor, which augments attribute sizes in its modes over time. Synchronized behavior tends to form dense blocks (i.e.~subtensors) in such a tensor, signaling anomalous behavior, or interesting communities. However, existing dense block detection methods are either based on a static tensor, or lack an efficient algorithm in a streaming setting. Therefore, we propose a fast streaming algorithm, AUGSPLICING, which can detect the top dense blocks by incrementally splicing the previous detection with the incoming ones in new tuples, avoiding re-runs over all the history data at every tracking time step. AUGSPLICING is based on a splicing condition that guides the algorithm (Section 4). Compared to the state-of-the-art methods, our method is (1) effective to detect fraudulent behavior in installing data of real-world apps and find a synchronized group of students with interesting features in campus Wi-Fi data; (2) robust with splicing theory for dense block detection; (3) streaming and faster than the existing streaming algorithm, with closely comparable accuracy.

TipTap: Approximate Mining of Frequent k -Subgraph Patterns in Evolving Graphs

“Perhaps he could dance first and think afterwards, if it isn’t too much to ask him.” S. Beckett, Waiting for Godot Given a labeled graph, the collection of k -vertex induced connected subgraph patterns that appear in the graph more frequently than a user-specified minimum threshold provides a compact summary of the characteristics of the graph, and finds applications ranging from biology to network science. However, finding these patterns is challenging, even more so for dynamic graphs that evolve over time, due to the streaming nature of the input and the exponential time complexity of the problem. We study this task in both incremental and fully-dynamic streaming settings, where arbitrary edges can be added or removed from the graph. We present TipTap , a suite of algorithms to compute high-quality approximations of the frequent k -vertex subgraphs w.r.t. a given threshold, at any time (i.e., point of the stream), with high probability. In contrast to existing state-of-the-art solutions that require iterating over the entire set of subgraphs in the vicinity of the updated edge, TipTap operates by efficiently maintaining a uniform sample of connected k -vertex subgraphs, thanks to an optimized neighborhood-exploration procedure. We provide a theoretical analysis of the proposed algorithms in terms of their unbiasedness and of the sample size needed to obtain a desired approximation quality. Our analysis relies on sample-complexity bounds that use Vapnik–Chervonenkis dimension, a key concept from statistical learning theory, which allows us to derive a sufficient sample size that is independent from the size of the graph. The results of our empirical evaluation demonstrates that TipTap returns high-quality results more efficiently and accurately than existing baselines.

Streaming Algorithms for Maximizing Monotone DR-Submodular Functions with a Cardinality Constraint on the Integer Lattice

Emerging applications such as optimal budget allocation and sensor placement impose problems of maximizing variants of submodular functions with constraints under a streaming setting. In this paper, we first devise a streaming algorithm based on Sieve-Streaming for maximizing a monotone diminishing return submodular (DR-submodular) function with a cardinality constraint on the integer lattice and show it is a one-pass algorithm with approximation ratio [Formula: see text]. The key idea to ensure one pass for the algorithm is to combine binary search for determining the level of an element with the exponential-growth method for estimating the OPT. Inspired by Sieve-Streaming++, we then improve the memory of the algorithm to [Formula: see text] and the query complexity to [Formula: see text].

Improved Streaming Algorithms for Maximizing Monotone Submodular Functions under a Knapsack Constraint

In this paper, we consider the problem of maximizing a monotone submodular function subject to a knapsack constraint in a streaming setting. In such a setting, elements arrive sequentially and at any point in time, and the algorithm can store only a small fraction of the elements that have arrived so far. For the special case that all elements have unit sizes (i.e., the cardinality-constraint case), one can find a \((0.5-\varepsilon )\)-approximate solution in \(O(K\varepsilon ^{-1})\) space, where K is the knapsack capacity (Badanidiyuru et al. KDD 2014). The approximation ratio is recently shown to be optimal (Feldman et al. STOC 2020). In this work, we propose a \((0.4-\varepsilon )\)-approximation algorithm for the knapsack-constrained problem, using space that is a polynomial of K and \(\varepsilon \). This improves on the previous best ratio of \(0.363-\varepsilon \) with space of the same order. Our algorithm is based on a careful combination of various ideas to transform multiple-pass streaming algorithms into a single-pass one.

Latent Dirichlet Allocation Model Training With Differential Privacy

Latent Dirichlet Allocation (LDA) is a popular topic modeling technique for hidden semantic discovery of text data and serves as a fundamental tool for text analysis in various applications. However, the LDA model as well as the training process of LDA may expose the text information in the training data, thus bringing significant privacy concerns. To address the privacy issue in LDA, we systematically investigate the privacy protection of the main-stream LDA training algorithm based on Collapsed Gibbs Sampling (CGS) and propose several differentially private LDA algorithms for typical training scenarios. In particular, we present the first theoretical analysis on the inherent differential privacy guarantee of CGS based LDA training and further propose a centralized privacy-preserving algorithm (HDP-LDA) that can prevent data inference from the intermediate statistics in the CGS training. Also, we propose a locally private LDA training algorithm (LP-LDA) on crowdsourced data to provide local differential privacy for individual data contributors. Furthermore, we extend LP-LDA to an online version as OLP-LDA to achieve LDA training on locally private mini-batches in a streaming setting. Extensive analysis and experiment results validate both the effectiveness and efficiency of our proposed privacy-preserving LDA training algorithms.

Spectral Algorithms for Streaming Graph Analysis: A Survey

Streaming data models refer to some constrained settings through which continuous flow of information regarding updates on the data becomes available. Graphs can also be represented in a streaming setting where interaction information turns out to be accessible as a stream of inclusion or exclusion of interactions. Analysis of streaming graphs helps to understand extreme-scale and dynamic real-life interactions in different forms. The growth of world wide web has drastically changed the way we look at various real-life evolving gigantic networks. This has motivated the development of streaming algorithms to be applied on graphs at scale. To achieve this scalability, sketching and sampling strategies are generally adopted to realize the different attributes of graphs. Spectrum of a graph, being one of the most appreciated characteristics, has lead to the evolution of an entire class of spectral algorithms. In this paper, we touch upon the state-of-the-art progress in streaming graph analysis with spectral algorithms. We mainly cover the latest developments in the areas like sampling, sparsification, singular value decomposition, counting problems related to local structures, analysis of global structures, partitioning, labeling, mesh processing, discovery of patterns, anomalous hotspot discovery, detection of communities, etc. on the subject of streaming graphs.

Biclustering and boolean matrix factorization in data streams

We study clustering of bipartite graphs and Boolean matrix factorization in data streams. We consider a streaming setting in which the vertices from the left side of the graph arrive one by one together with all of their incident edges. We provide an algorithm which after one pass over the stream recovers the set of clusters on the right side of the graph using sublinear space; to the best of our knowledge this is the first algorithm with this property. We also show that after a second pass over the stream the left clusters of the bipartite graph can be recovered and we show how to extend our algorithm to solve the Boolean matrix factorization problem (by exploiting the correspondence of Boolean matrices and bipartite graphs). We evaluate an implementation of the algorithm on synthetic data and on real-world data. On real-world datasets the algorithm is orders of magnitudes faster than a static baseline algorithm while providing quality results within a factor 2 of the baseline algorithm. Our algorithm scales linearly in the number of edges in the graph. Finally, we analyze the algorithm theoretically and provide sufficient conditions under which the algorithm recovers a set of planted clusters under a standard random graph model.

Similarity Join and Similarity Self-Join Size Estimation in a Streaming Environment

We study the problem of similarity self-join and similarity join size estimation in a streaming setting where the goal is to estimate, in one scan of the input and with sublinear space in the input size, the number of record pairs that have a similarity within a given threshold. The problem has many applications in data cleaning and query plan generation, where the cost of a similarity join may be estimated before actually doing the join. On unary input where two records either match or don't match, the problem becomes join and self-join size estimation for which one-pass algorithms are readily available. Our work addresses the problem for d-ary input, for d >= 1, where the degree of similarity can vary from 1 to d. We show that our proposed algorithm gives an accurate estimate and scales well with the input size. We provide error bounds and time and space costs, and conduct an extensive experimental evaluation of our algorithm, comparing its estimation accuracy to a few competitors, including some multi-pass algorithms. Our results show that given the same space, the proposed algorithm has an order of magnitude less error for a large range of similarity thresholds.

Estimation of Gaussian mixture models via tensor moments with application to online learning

In this paper, we present an alternating gradient descent algorithm for estimating parameters of a spherical Gaussian mixture model by the method of moments (AGD-MoM). We formulate the problem as a constrained optimisation problem which simultaneously matches the third order moments from the data, represented as a tensor, and the second order moment, which is the empirical covariance matrix. We derive the necessary gradients (and second derivatives), and use them to implement alternating gradient search to estimate the parameters of the model. We show that the proposed method is applicable in both a batch as well as in a streaming (online) setting. Using synthetic and benchmark datasets, we demonstrate empirically that the proposed algorithm outperforms the more classical algorithms like Expectation Maximisation and variational Bayes.

Monitoring high-frequency data streams in FinTech: FADO versus K-means

Modern applications of FinTech are challenged by enormous volumes of financial data. One way to handle these is to adopt a streaming setting where data are only available to the algorithms during a very short time. When a new data point (financial transaction) is generated, it needs to be processed directly, and be forgotten immediately after. Especially, ongoing globalization efforts in FinTech require modern methods of fault detection to be able to work efficiently through more than 10 000 financial transactions per second if they are to be deployed as a first line of defence. This article investigates two algorithms able to perform well in this demanding setting: $K$K-means and FADO. Especially, this article provides supports for the claim that “the use of multiple clusters does not necessarily translate into increased detection performance.” To support this claim, results are reported when operating in a quasi-realistic case study of Anti Money Laundering (AML) detection in real-time payment systems. We focus on two prototypical algorithms: the passive aggressive FADO assuming a single cluster, and the well-known $K$K-means algorithm working with $K>1$K>1 clusters. We find—in this case—that the use of $K$K-means with multiple clusters is unfavorable as 1) both tuning for $K$K, as well as the need for additional complexity in the $K$K-means algorithm challenges the computational constraints; 2) $K$K-means introduces necessarily added variability (unreliability) in the results; 3) it requires dimensionality reduction, compromising interpretability of the detections; 4) the prevalence of singleton clusters adds unreliability to the outcome. This makes in the presented case FADO favorable over K-means (with $K>1$K>1).

Streaming Algorithms for Maximizing Monotone Submodular Functions Under a Knapsack Constraint

In this paper, we consider the problem of maximizing a monotone submodular function subject to a knapsack constraint in a streaming setting. In such a setting, elements arrive sequentially and at any point in time, and the algorithm can store only a small fraction of the elements that have arrived so far. For the special case that all elements have unit sizes (i.e., the cardinality-constraint case), one can find a $$(0.5-\varepsilon )$$ -approximate solution in $$O(K\varepsilon ^{-1})$$ space, where K is the knapsack capacity (Badanidiyuru et al. KDD 2014). The approximation ratio is recently shown to be optimal (Feldman et al. STOC 2020). In this work, we propose a $$(0.4-\varepsilon )$$ -approximation algorithm for the knapsack-constrained problem, using space that is a polynomial of K and $$\varepsilon $$ . This improves on the previous best ratio of $$0.363-\varepsilon $$ with space of the same order. Our algorithm is based on a careful combination of various ideas to transform multiple-pass streaming algorithms into a single-pass one.

Real-time distributed co-movement pattern detection on streaming trajectories

With the widespread deployment of mobile devices with positioning capabilities, increasingly massive volumes of trajectory data are being collected that capture the movements of people and vehicles. This data enables co-movement pattern detection, which is important in applications such as trajectory compression and future-movement prediction. Existing co-movement pattern detection studies generally consider historical data and thus propose offline algorithms. However, applications such as future movement prediction need real-time processing over streaming trajectories. Thus, we investigate real-time distributed co-movement pattern detection over streaming trajectories. Existing off-line methods assume that all data is available when the processing starts. Nevertheless, in a streaming setting, unbounded data arrives in real time, making pattern detection challenging. To this end, we propose a framework based on Apache Flink, which is designed for efficient distributed streaming data processing. The framework encompasses two phases: clustering and pattern enumeration. To accelerate the clustering, we use a range join based on two-layer indexing, and provide techniques that eliminate unnecessary verifications. To perform pattern enumeration efficiently, we present two methods FBA and VBA that utilize id-based partitioning. When coupled with bit compression and candidate-based enumeration techniques, we reduce the enumeration cost from exponential to linear. Extensive experiments offer insight into the efficiency of the proposed framework and its constituent techniques compared with existing methods.

User guidance for efficient fact checking

The Web constitutes a valuable source of information. In recent years, it fostered the construction of large-scale knowledge bases, such as Freebase, YAGO, and DBpedia. The open nature of the Web, with content potentially being generated by everyone, however, leads to inaccuracies and misinformation. Construction and maintenance of a knowledge base thus has to rely on fact checking, an assessment of the credibility of facts. Due to an inherent lack of ground truth information, such fact checking cannot be done in a purely automated manner, but requires human involvement. In this paper, we propose a comprehensive framework to guide users in the validation of facts, striving for a minimisation of the invested effort. Our framework is grounded in a novel probabilistic model that combines user input with automated credibility inference. Based thereon, we show how to guide users in fact checking by identifying the facts for which validation is most beneficial. Moreover, our framework includes techniques to reduce the manual effort invested in fact checking by determining when to stop the validation and by supporting efficient batching strategies. We further show how to handle fact checking in a streaming setting. Our experiments with three real-world datasets demonstrate the efficiency and effectiveness of our framework: A knowledge base of high quality, with a precision of above 90%, is constructed with only a half of the validation effort required by baseline techniques.

Hierarchical Partial Planarity

In this paper we consider graphs whose edges are associated with a degree of {\em importance}, which may depend on the type of connections they represent or on how recently they appeared in the scene, in a streaming setting. The goal is to construct layouts of these graphs in which the readability of an edge is proportional to its importance, that is, more important edges have fewer crossings. We formalize this problem and study the case in which there exist three different degrees of importance. We give a polynomial-time testing algorithm when the graph induced by the two most important sets of edges is biconnected. We also discuss interesting relationships with other constrained-planarity problems.