arXiv subjects
Publications and source records attributed to Lior Rokach.
In this work we implement a training of a Language Model (LM), using Recurrent Neural Network (RNN) and GloVe word embeddings, introduced by Pennigton et al. in [1]. The implementation is following the general idea of training RNNs for LM tasks presented in [2], but is rather using Gated Recurrent Unit (GRU) [3] for a memory cell, and not the more commonly used LSTM [4].
Driving under the influence of alcohol is a widespread phenomenon in the US where it is considered a major cause of fatal accidents. In this research we present a novel approach and concept for detecting intoxication from motion differences obtained by the sensors of wearable devices. We formalize the problem of drunkenness detection as a supervised machine learning task, both as a binary classification problem (drunk or sober) and a regression problem (the breath alcohol content level). In order to test our approach, we collected data from 30 different subjects (patrons at three bars) using Google Glass and the LG G-watch, Microsoft Band, and Samsung Galaxy S4. We validated our results against an admissible breathalyzer used by the police. A system based on this concept, successfully detected intoxication and achieved the following results: 0.95 AUC and 0.05 FPR, given a fixed TPR of 1.0. Applications based on our system can be used to analyze the free gait of drinkers when they walk from the car to the bar and vice-versa, in order to alert people, or even a connected car and prevent people from driving under the influence of alcohol.
We present a new concept - Wikiometrics - the derivation of metrics and indicators from Wikipedia. Wikipedia provides an accurate representation of the real world due to its size, structure, editing policy and popularity. We demonstrate an innovative mining methodology, where different elements of Wikipedia - content, structure, editorial actions and reader reviews - are used to rank items in a manner which is by no means inferior to rankings produced by experts or other methods. We test our proposed method by applying it to two real-world ranking problems: top world universities and academic journals. Our proposed ranking methods were compared to leading and widely accepted benchmarks, and were found to be extremely correlative but with the advantage of the data being publically available.
Online Social Networks (OSNs), such as Facebook and Twitter, have become an integral part of our daily lives. There are hundreds of OSNs, each with its own focus in that each offers particular services and functionalities. Recent studies show that many OSN users create several accounts on multiple OSNs using the same or different personal information. Collecting all the available data of an individual from several OSNs and fusing it into a single profile can be useful for many purposes. In this paper, we introduce novel machine learning based methods for solving Entity Resolution (ER), a problem for matching user profiles across multiple OSNs. The presented methods are able to match between two user profiles from two different OSNs based on supervised learning techniques, which use features extracted from each one of the user profiles. By using the extracted features and supervised learning techniques, we developed classifiers which can perform entity matching between two profiles for the following scenarios: (a) matching entities across two OSNs; (b) searching for a user by similar name; and (c) de-anonymizing a user's identity. The constructed classifiers were tested by using data collected from two popular OSNs, Facebook and Xing. We then evaluated the classifiers' performances using various evaluation measures, such as true and false positive rates, accuracy, and the Area Under the receiver operator Curve (AUC). The constructed classifiers were evaluated and their classification performance measured by AUC was quite remarkable, with an AUC of up to 0.982 and an accuracy of up to 95.9% in identifying user profiles across two OSNs.
Each new identified security threat class triggers new research and development efforts by the scientific and professional communities. In this study, we investigate the rate at which the scientific and professional communities react to new identified threat classes as it is reflected in the number of patents, scientific articles and professional publications over a long period of time. The following threat classes were studied: Phishing; SQL Injection; BotNet; Distributed Denial of Service; and Advanced Persistent Threat. Our findings suggest that in most cases it takes a year for the scientific community and more than two years for industry to react to a new threat class with patents. Since new products follow patents, it is reasonable to expect that there will be a window of approximately two to three years in which no effective product is available to cope with the new threat class.
Selecting the best classifier among the available ones is a difficult task, especially when only instances of one class exist. In this work we examine the notion of combining one-class classifiers as an alternative for selecting the best classifier. In particular, we propose two new one-class classification performance measures to weigh classifiers and show that a simple ensemble that implements these measures can outperform the most popular one-class ensembles. Furthermore, we propose a new one-class ensemble scheme, TUPSO, which uses meta-learning to combine one-class classifiers. Our experiments demonstrate the superiority of TUPSO over all other tested ensembles and show that the TUPSO performance is statistically indistinguishable from that of the hypothetical best classifier.
Ensemble methods have been shown to be an effective tool for solving multi-label classification tasks. In the RAndom k-labELsets (RAKEL) algorithm, each member of the ensemble is associated with a small randomly-selected subset of k labels. Then, a single label classifier is trained according to each combination of elements in the subset. In this paper we adopt a similar approach, however, instead of randomly choosing subsets, we select the minimum required subsets of k labels that cover all labels and meet additional constraints such as coverage of inter-label correlations. Construction of the cover is achieved by formulating the subset selection as a minimum set covering problem (SCP) and solving it by using approximation algorithms. Every cover needs only to be prepared once by offline algorithms. Once prepared, a cover may be applied to the classification of any given multi-label dataset whose properties conform with those of the cover. The contribution of this paper is two-fold. First, we introduce SCP as a general framework for constructing label covers while allowing the user to incorporate cover construction constraints. We demonstrate the effectiveness of this framework by proposing two construction constraints whose enforcement produces covers that improve the prediction performance of random selection. Second, we provide theoretical bounds that quantify the probabilities of random selection to produce covers that meet the proposed construction criteria. The experimental results indicate that the proposed methods improve multi-label classification accuracy and stability compared with the RAKEL algorithm and to other state-of-the-art algorithms.
XML transactions are used in many information systems to store data and interact with other systems. Abnormal transactions, the result of either an on-going cyber attack or the actions of a benign user, can potentially harm the interacting systems and therefore they are regarded as a threat. In this paper we address the problem of anomaly detection and localization in XML transactions using machine learning techniques. We present a new XML anomaly detection framework, XML-AD. Within this framework, an automatic method for extracting features from XML transactions was developed as well as a practical method for transforming XML features into vectors of fixed dimensionality. With these two methods in place, the XML-AD framework makes it possible to utilize general learning algorithms for anomaly detection. Central to the functioning of the framework is a novel multi-univariate anomaly detection algorithm, ADIFA. The framework was evaluated on four XML transactions datasets, captured from real information systems, in which it achieved over 89% true positive detection rate with less than a 0.2% false positive rate.
We present a novel approach for the construction of ensemble classifiers based on dimensionality reduction. Dimensionality reduction methods represent datasets using a small number of attributes while preserving the information conveyed by the original dataset. The ensemble members are trained based on dimension-reduced versions of the training set. These versions are obtained by applying dimensionality reduction to the original training set using different values of the input parameters. This construction meets both the diversity and accuracy criteria which are required to construct an ensemble classifier where the former criterion is obtained by the various input parameter values and the latter is achieved due to the decorrelation and noise reduction properties of dimensionality reduction. In order to classify a test sample, it is first embedded into the dimension reduced space of each individual classifier by using an out-of-sample extension algorithm. Each classifier is then applied to the embedded sample and the classification is obtained via a voting scheme. We present three variations of the proposed approach based on the Random Projections, the Diffusion Maps and the Random Subspaces dimensionality reduction algorithms. We also present a multi-strategy ensemble which combines AdaBoost and Diffusion Maps. A comparison is made with the Bagging, AdaBoost, Rotation Forest ensemble classifiers and also with the base classifier which does not incorporate dimensionality reduction. Our experiments used seventeen benchmark datasets from the UCI repository. The results obtained by the proposed algorithms were superior in many cases to other algorithms.
In this paper we present a new approach to content-based transfer learning for solving the data sparsity problem in cases when the users' preferences in the target domain are either scarce or unavailable, but the necessary information on the preferences exists in another domain. We show that training a system to use such information across domains can produce better performance. Specifically, we represent users' behavior patterns based on topological graph structures. Each behavior pattern represents the behavior of a set of users, when the users' behavior is defined as the items they rated and the items' rating values. In the next step we find a correlation between behavior patterns in the source domain and behavior patterns in the target domain. This mapping is considered a bridge between the two domains. Based on the correlation and content-attributes of the items, we train a machine learning model to predict users' ratings in the target domain. When we compare our approach to the popularity approach and KNN-cross-domain on a real world dataset, the results show that on an average of 83$%$ of the cases our approach outperforms both methods.
Our system illustrates how information retrieved from social networks can be used for suggesting experts for specific tasks. The system is designed to facilitate the task of finding the appropriate person(s) for a job, as a conference committee member, an advisor, etc. This short description will demonstrate how the system works in the context of the HCIR2012 published tasks.
Singular Value Decomposition (SVD) has been used successfully in recent years in the area of recommender systems. In this paper we present how this model can be extended to consider both user ratings and information from Wikipedia. By mapping items to Wikipedia pages and quantifying their similarity, we are able to use this information in order to improve recommendation accuracy, especially when the sparsity is high. Another advantage of the proposed approach is the fact that it can be easily integrated into any other SVD implementation, regardless of additional parameters that may have been added to it. Preliminary experimental results on the MovieLens dataset are encouraging.
In this paper we examine the effect of applying ensemble learning to the performance of collaborative filtering methods. We present several systematic approaches for generating an ensemble of collaborative filtering models based on a single collaborative filtering algorithm (single-model or homogeneous ensemble). We present an adaptation of several popular ensemble techniques in machine learning for the collaborative filtering domain, including bagging, boosting, fusion and randomness injection. We evaluate the proposed approach on several types of collaborative filtering base models: k- NN, matrix factorization and a neighborhood matrix factorization model. Empirical evaluation shows a prediction improvement compared to all base CF algorithms. In particular, we show that the performance of an ensemble of simple (weak) CF models such as k-NN is competitive compared with a single strong CF model (such as matrix factorization) while requiring an order of magnitude less computational cost.
Time series classification is a field which has drawn much attention over the past decade. A new approach for classification of time series uses classification trees based on shapelets. A shapelet is a subsequence extracted from one of the time series in the dataset. A disadvantage of this approach is the time required for building the shapelet-based classification tree. The search for the best shapelet requires examining all subsequences of all lengths from all time series in the training set. A key goal of this work was to find an evaluation order of the shapelets space which enables fast convergence to an accurate model. The comparative analysis we conducted clearly indicates that a random evaluation order yields the best results. Our empirical analysis of the distribution of high-quality shapelets within the shapelets space provides insights into why randomized shapelets sampling is superior to alternative evaluation orders. We present an algorithm for randomized model generation for shapelet-based classification that converges extremely quickly to a model with surprisingly high accuracy after evaluating only an exceedingly small fraction of the shapelets space.