Author name cluster

Doina Caragea

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

5 papers

1 author row

AAAI Conference 2020 Conference Paper

On Identifying Hashtags in Disaster Twitter Data

Jishnu Ray Chowdhury
Cornelia Caragea
Doina Caragea

Tweet hashtags have the potential to improve the search for information during disaster events. However, there is a large number of disaster-related tweets that do not have any user-provided hashtags. Moreover, only a small number of tweets that contain actionable hashtags are useful for disaster response. To facilitate progress on automatic identiﬁcation (or extraction) of disaster hashtags for Twitter data, we construct a unique dataset of disaster-related tweets annotated with hashtags useful for ﬁltering actionable information. Using this dataset, we further investigate Long Short- Term Memory-based models within a Multi-Task Learning framework. The best performing model achieves an F1-score as high as 92. 22%. The dataset, code, and other resources are available on Github. 1

PDF Details

AAAI Conference 2019 Short Paper

Identifying Android Malware Using Network-Based Approaches

Emily Alfs
Doina Caragea
Nathan Albin
Pietro Poggi-Corradini

The proliferation of Android apps has resulted in many malicious apps entering the market and causing significant damage. Robust techniques that determine if an app is malicious are greatly needed. We propose the use of a network-based approach to effectively separate malicious from benign apps, based on a small labeled dataset. The apps in our dataset come from the Google Play Store and have been scanned for malicious behavior using Virus Total to produce a ground truth dataset with labels malicous or benign. The apps in the resulting dataset have been represented using binary feature vectors (where the features represent permissions, intent actions, discriminative APIs, obfuscation signatures, and native code signatures). We have used the feature vectors corresponding to apps to build a weighted network that captures the “closeness” between apps. We propagate labels from the labeled apps to unlabeled apps, and evaluate the effectiveness of the proposed approach using the F1-measure. We have conducted experiments to compare three variants of the label propagation approaches on datasets that include increasingly larger amounts of labeled data. The results have shown that a variant proposed in this study gives the best results overall.

PDF Details

AAAI Conference 2017 Short Paper

Android Malware Detection with Weak Ground Truth Data

Jordan DeLoach
Doina Caragea
Xinming Ou

For Android malware detection, precise ground truth is a rare commodity. As security knowledge evolves, what may be considered ground truth at one moment in time may change, and apps once considered benign may turn out to be malicious. The inevitable noise in data labels poses a challenge to inferring effective machine learning classiﬁers. Our work is focused on approaches for learning classiﬁers for Android malware detection in a manner that is methodologically sound with regard to the uncertain and ever-changing ground truth in the problem space. We leverage the fact that although data labels are unavoidably noisy, a malware label is much more precise than a benign label. While you can be conﬁdent that an app is malicious, you can never be certain that a benign app is really benign, or just undetected malware. Based on this insight, we leverage a modiﬁed Logistic Regression classiﬁer that allows us to learn from only positive and unlabeled data, without making any assumptions about benign labels. We ﬁnd Label Regularized Logistic Regression to perform well for noisy app datasets, as well as datasets where there is a limited amount of positive labeled data, both of which are representative of real-world situations.

PDF Details

AAAI Conference 2002 Short Paper

Learning in Open-Ended Dynamic Distributed Environments

Doina Caragea

In some domains (e.g., molecular biology), data repositories are large in size, dynamic, and physically distributed. Consequently, it is neither desirable nor feasible to gather all the data in a centralized location for analysis. Hence, efficient distributed learning algorithms that can operate across multiple data sources without the need to transmit large amounts of data and cumulative learning algorithms that can cope with data sets that grow at rapid rate are needed. We formulate a class of distributed and cumulative learning problems, and present a general strategy for transforming a large class of traditional machine learning algorithms into distributed and cumulative learning algorithms. Our general strategy is based on a decomposition of the learning task into information extraction and hypothesis generation components. We use this approach to construct provably exact distributed algorithms for support vector machines and also for decision tree learning. We formalize the treatment of distributed learning by introducing a family of learning, information extraction and information composition operators and establishing sufficient conditions for provably exact distributed and cumulative learning in terms of general algebraic properties of the operators.

PDF Details

AAAI Conference 2000 Short Paper

Incremental and Distributed Learning with Support Vector Machines

Doina Caragea
and Vasant Honavar

Due to the increase in the amount of data gathered every day in the real world problems (e.g., bioinformatics), there is a need for inductive learning algorithms that can incrementally process large amounts of data that is being accumulated over time in physically distributed, autonomous data repositories. In the incremental setting, the learner gradually refines a hypothesis (or a set of hypotheses) as new data become available. Because of the large volume of data involved, it may not be practical to store and access the entire dataset during learning. Thus, the learner does not have access to data that has been encountered at a previous time. Learning in the distributed setting can be defined in a similar fashion. An incremental or distributed learning algorithm is said to be exact if it gives the same results as those obtained by batch learning (i.e., when the entire dataset is accessible to the learning algorithm during learning). We explore exact distributed and incremental learning algorithms that are variants and extensions of the support vector machine (SVM) family of learning algorithms.

PDF Details