Niklas Kolbe, P. Vandenbussche, S. Kubler, Yves Le Traon
Ontology search and ranking are key building blocks to establish and reuse shared conceptualizations of domain knowledge on the Web. However, the effectiveness of proposed ontology ranking models is difficult to compare since these are often evaluated on diverse datasets that are limited by their static nature and scale. In this paper, we first introduce the LOVBench dataset as a benchmark for ontology term ranking. With inferred relevance judgments for more than 7000 queries, LOVBench is large enough to perform a comparison study using learning to rank (LTR) with complex ontology ranking models. Instead of relying on relevance judgments from a few experts, we consider implicit feedback from many actual users collected from the Linked Open Vocabularies (LOV) platform. Our approach further enables continuous updates of the benchmark, capturing the evolution of ontologies’ relevance in an ever-changing data community. Second, we compare the performance of several feature configurations from the literature using LOVBench in LTR settings and discuss the results in the context of the observed real-world user behavior. Our experimental results show that feature configurations which are (i) well-suited to the user behavior, (ii) cover all features types, and (iii) consider decomposition of features can significantly improve the ranking performance.
{"title":"LOVBench: Ontology Ranking Benchmark","authors":"Niklas Kolbe, P. Vandenbussche, S. Kubler, Yves Le Traon","doi":"10.1145/3366423.3380245","DOIUrl":"https://doi.org/10.1145/3366423.3380245","url":null,"abstract":"Ontology search and ranking are key building blocks to establish and reuse shared conceptualizations of domain knowledge on the Web. However, the effectiveness of proposed ontology ranking models is difficult to compare since these are often evaluated on diverse datasets that are limited by their static nature and scale. In this paper, we first introduce the LOVBench dataset as a benchmark for ontology term ranking. With inferred relevance judgments for more than 7000 queries, LOVBench is large enough to perform a comparison study using learning to rank (LTR) with complex ontology ranking models. Instead of relying on relevance judgments from a few experts, we consider implicit feedback from many actual users collected from the Linked Open Vocabularies (LOV) platform. Our approach further enables continuous updates of the benchmark, capturing the evolution of ontologies’ relevance in an ever-changing data community. Second, we compare the performance of several feature configurations from the literature using LOVBench in LTR settings and discuss the results in the context of the observed real-world user behavior. Our experimental results show that feature configurations which are (i) well-suited to the user behavior, (ii) cover all features types, and (iii) consider decomposition of features can significantly improve the ranking performance.","PeriodicalId":20754,"journal":{"name":"Proceedings of The Web Conference 2020","volume":"40 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2020-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89604931","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In the modern tourism industry, next point-of-interest (POI) recommendation is an important mobile service as it effectively aids hesitating travelers to decide the next POI to visit. Currently, most next POI recommender systems are built upon a cloud-based paradigm, where the recommendation models are trained and deployed on the powerful cloud servers. When a recommendation request is made by a user via mobile devices, the current contextual information will be uploaded to the cloud servers to help the well-trained models generate personalized recommendation results. However, in reality, this paradigm heavily relies on high-quality network connectivity, and is subject to high energy footprint in the operation and increasing privacy concerns among the public. To bypass these defects, we propose a novel Light Location Recommender System (LLRec) to perform next POI recommendation locally on resource-constrained mobile devices. To make LLRec fully compatible with the limited computing resources and memory space, we leverage FastGRNN, a lightweight but effective gated Recurrent Neural Network (RNN) as its main building block, and significantly compress the model size by adopting the tensor-train composition in the embedding layer. As a compact model, LLRec maintains its robustness via an innovative teacher-student training framework, where a powerful teacher model is trained on the cloud to learn essential knowledge from available contextual data, and the simplified student model LLRec is trained under the guidance of the teacher model. The final LLRec is downloaded and deployed on users’ mobile devices to generate accurate recommendations solely utilizing users’ local data. As a result, LLRec significantly reduces the dependency on cloud servers, thus allowing for next POI recommendation in a stable, cost-effective and secure way. Extensive experiments on two large-scale recommendation datasets further demonstrate the superiority of our proposed solution.
{"title":"Next Point-of-Interest Recommendation on Resource-Constrained Mobile Devices","authors":"Qinyong Wang, Hongzhi Yin, Tong Chen, Zi Huang, Hao Wang, Yanchang Zhao, Nguyen Quoc Viet Hung","doi":"10.1145/3366423.3380170","DOIUrl":"https://doi.org/10.1145/3366423.3380170","url":null,"abstract":"In the modern tourism industry, next point-of-interest (POI) recommendation is an important mobile service as it effectively aids hesitating travelers to decide the next POI to visit. Currently, most next POI recommender systems are built upon a cloud-based paradigm, where the recommendation models are trained and deployed on the powerful cloud servers. When a recommendation request is made by a user via mobile devices, the current contextual information will be uploaded to the cloud servers to help the well-trained models generate personalized recommendation results. However, in reality, this paradigm heavily relies on high-quality network connectivity, and is subject to high energy footprint in the operation and increasing privacy concerns among the public. To bypass these defects, we propose a novel Light Location Recommender System (LLRec) to perform next POI recommendation locally on resource-constrained mobile devices. To make LLRec fully compatible with the limited computing resources and memory space, we leverage FastGRNN, a lightweight but effective gated Recurrent Neural Network (RNN) as its main building block, and significantly compress the model size by adopting the tensor-train composition in the embedding layer. As a compact model, LLRec maintains its robustness via an innovative teacher-student training framework, where a powerful teacher model is trained on the cloud to learn essential knowledge from available contextual data, and the simplified student model LLRec is trained under the guidance of the teacher model. The final LLRec is downloaded and deployed on users’ mobile devices to generate accurate recommendations solely utilizing users’ local data. As a result, LLRec significantly reduces the dependency on cloud servers, thus allowing for next POI recommendation in a stable, cost-effective and secure way. Extensive experiments on two large-scale recommendation datasets further demonstrate the superiority of our proposed solution.","PeriodicalId":20754,"journal":{"name":"Proceedings of The Web Conference 2020","volume":"83 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2020-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82371558","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
David Zeber, Sarah Bird, Camila Oliveira, Walter Rudametkin, I. Segall, Fredrik Wollsén, M. Lopatka
Large-scale Web crawls have emerged as the state of the art for studying characteristics of the Web. In particular, they are a core tool for online tracking research. Web crawling is an attractive approach to data collection, as crawls can be run at relatively low infrastructure cost and don’t require handling sensitive user data such as browsing histories. However, the biases introduced by using crawls as a proxy for human browsing data have not been well studied. Crawls may fail to capture the diversity of user environments, and the snapshot view of the Web presented by one-time crawls does not reflect its constantly evolving nature, which hinders reproducibility of crawl-based studies. In this paper, we quantify the repeatability and representativeness of Web crawls in terms of common tracking and fingerprinting metrics, considering both variation across crawls and divergence from human browser usage. We quantify baseline variation of simultaneous crawls, then isolate the effects of time, cloud IP address vs. residential, and operating system. This provides a foundation to assess the agreement between crawls visiting a standard list of high-traffic websites and actual browsing behaviour measured from an opt-in sample of over 50,000 users of the Firefox Web browser. Our analysis reveals differences between the treatment of stateless crawling infrastructure and generally stateful human browsing, showing, for example, that crawlers tend to experience higher rates of third-party activity than human browser users on loading pages from the same domains.
{"title":"The Representativeness of Automated Web Crawls as a Surrogate for Human Browsing","authors":"David Zeber, Sarah Bird, Camila Oliveira, Walter Rudametkin, I. Segall, Fredrik Wollsén, M. Lopatka","doi":"10.1145/3366423.3380104","DOIUrl":"https://doi.org/10.1145/3366423.3380104","url":null,"abstract":"Large-scale Web crawls have emerged as the state of the art for studying characteristics of the Web. In particular, they are a core tool for online tracking research. Web crawling is an attractive approach to data collection, as crawls can be run at relatively low infrastructure cost and don’t require handling sensitive user data such as browsing histories. However, the biases introduced by using crawls as a proxy for human browsing data have not been well studied. Crawls may fail to capture the diversity of user environments, and the snapshot view of the Web presented by one-time crawls does not reflect its constantly evolving nature, which hinders reproducibility of crawl-based studies. In this paper, we quantify the repeatability and representativeness of Web crawls in terms of common tracking and fingerprinting metrics, considering both variation across crawls and divergence from human browser usage. We quantify baseline variation of simultaneous crawls, then isolate the effects of time, cloud IP address vs. residential, and operating system. This provides a foundation to assess the agreement between crawls visiting a standard list of high-traffic websites and actual browsing behaviour measured from an opt-in sample of over 50,000 users of the Firefox Web browser. Our analysis reveals differences between the treatment of stateless crawling infrastructure and generally stateful human browsing, showing, for example, that crawlers tend to experience higher rates of third-party activity than human browser users on loading pages from the same domains.","PeriodicalId":20754,"journal":{"name":"Proceedings of The Web Conference 2020","volume":"46 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2020-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80370832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Although pronounceability can improve password memorability, most existing password generation approaches have not properly integrated the pronounceability of passwords in their designs. In this work, we demonstrate several shortfalls of current pronounceable password generation approaches, and then propose, ProSemPass, a new method of generating passwords that are pronounceable and semantically meaningful. In our approach, users supply initial input words and our system improves the pronounceability and meaning of the user-provided words by automatically creating a portmanteau. To measure the strength of our approach, we use attacker models, where attackers have complete knowledge of our password generation algorithms. We measure strength in guess numbers and compare those with other existing password generation approaches. Using a large-scale IRB-approved user study with 1,563 Amazon MTurkers over 9 different conditions, our approach achieves a 30% higher recall than those from current pronounceable password approaches, and is stronger than the offline guessing attack limit.
{"title":"How Do We Create a Fantabulous Password?","authors":"Simon S. Woo","doi":"10.1145/3366423.3380222","DOIUrl":"https://doi.org/10.1145/3366423.3380222","url":null,"abstract":"Although pronounceability can improve password memorability, most existing password generation approaches have not properly integrated the pronounceability of passwords in their designs. In this work, we demonstrate several shortfalls of current pronounceable password generation approaches, and then propose, ProSemPass, a new method of generating passwords that are pronounceable and semantically meaningful. In our approach, users supply initial input words and our system improves the pronounceability and meaning of the user-provided words by automatically creating a portmanteau. To measure the strength of our approach, we use attacker models, where attackers have complete knowledge of our password generation algorithms. We measure strength in guess numbers and compare those with other existing password generation approaches. Using a large-scale IRB-approved user study with 1,563 Amazon MTurkers over 9 different conditions, our approach achieves a 30% higher recall than those from current pronounceable password approaches, and is stronger than the offline guessing attack limit.","PeriodicalId":20754,"journal":{"name":"Proceedings of The Web Conference 2020","volume":"12 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2020-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74272261","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Distributed graph processing frameworks formulate tasks as sequences of supersteps within which communication is performed asynchronously by sending messages over the graph edges. PageRank’s communication pattern is identical across all its supersteps since each vertex sends messages to all its edges. We exploit this pattern to develop a new communication paradigm that allows us to exchange messages that include only edge payloads, dramatically reducing bandwidth requirements. Experiments on a web graph of 38 billion vertices and 3.1 trillion edges yield execution times of 34.4 seconds per iteration, suggesting more than an order of magnitude improvement over the state-of-the-art.
{"title":"Scaling PageRank to 100 Billion Pages","authors":"S. Stergiou","doi":"10.1145/3366423.3380035","DOIUrl":"https://doi.org/10.1145/3366423.3380035","url":null,"abstract":"Distributed graph processing frameworks formulate tasks as sequences of supersteps within which communication is performed asynchronously by sending messages over the graph edges. PageRank’s communication pattern is identical across all its supersteps since each vertex sends messages to all its edges. We exploit this pattern to develop a new communication paradigm that allows us to exchange messages that include only edge payloads, dramatically reducing bandwidth requirements. Experiments on a web graph of 38 billion vertices and 3.1 trillion edges yield execution times of 34.4 seconds per iteration, suggesting more than an order of magnitude improvement over the state-of-the-art.","PeriodicalId":20754,"journal":{"name":"Proceedings of The Web Conference 2020","volume":"34 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2020-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82794592","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Product image search in E-commerce systems is a challenging task, because of a huge number of product classes, low intra-class similarity and high inter-class similarity. Deep metric learning, based on paired distances independent of the number of classes, aims to minimize intra-class variances and inter-class similarity in feature embedding space. Most existing approaches strictly restrict the distance between samples with fixed values to distinguish different classes of samples. However, the distance of paired samples has various magnitudes during different training stages. Therefore, it is difficult to directly restrict absolute distances with fixed values. In this paper, we propose a novel Equidistant and Equidistributed Triplet-based (EET) loss function to adjust the distance between samples with relative distance constraints. By optimizing the loss function, the algorithm progressively maximizes intra-class similarity and inter-class variances. Specifically, 1) the equidistant loss pulls the matched samples closer by adaptively constraining two samples of the same class to be equally distant from another one of a different class in each triplet, 2) the equidistributed loss pushes the mismatched samples farther away by guiding different classes to be uniformly distributed while keeping intra-class structure compact in embedding space. Extensive experimental results on product search benchmarks verify the improved performance of our method. We also achieve improvements on other retrieval datasets, which show superior generalization capacity of our method in image search.
{"title":"Metric Learning with Equidistant and Equidistributed Triplet-based Loss for Product Image Search","authors":"Furong Xu, Wei Zhang, Yuan Cheng, Wei Chu","doi":"10.1145/3366423.3380094","DOIUrl":"https://doi.org/10.1145/3366423.3380094","url":null,"abstract":"Product image search in E-commerce systems is a challenging task, because of a huge number of product classes, low intra-class similarity and high inter-class similarity. Deep metric learning, based on paired distances independent of the number of classes, aims to minimize intra-class variances and inter-class similarity in feature embedding space. Most existing approaches strictly restrict the distance between samples with fixed values to distinguish different classes of samples. However, the distance of paired samples has various magnitudes during different training stages. Therefore, it is difficult to directly restrict absolute distances with fixed values. In this paper, we propose a novel Equidistant and Equidistributed Triplet-based (EET) loss function to adjust the distance between samples with relative distance constraints. By optimizing the loss function, the algorithm progressively maximizes intra-class similarity and inter-class variances. Specifically, 1) the equidistant loss pulls the matched samples closer by adaptively constraining two samples of the same class to be equally distant from another one of a different class in each triplet, 2) the equidistributed loss pushes the mismatched samples farther away by guiding different classes to be uniformly distributed while keeping intra-class structure compact in embedding space. Extensive experimental results on product search benchmarks verify the improved performance of our method. We also achieve improvements on other retrieval datasets, which show superior generalization capacity of our method in image search.","PeriodicalId":20754,"journal":{"name":"Proceedings of The Web Conference 2020","volume":"18 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2020-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79952749","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Shuguang Han, Michael Bendersky, Przemek Gajda, Sergey Novikov, Marc Najork, Bernhard Brodowsky, Alexandrin Popescul
The rapid growth of commercial web content has driven the development of shopping search services to help users find product offers. Due to the dynamic nature of commercial content, an effective recrawl policy is a key component in a shopping search service; it ensures that users have access to the up-to-date product details. Most of the existing strategies either relied on simple heuristics, or overlooked the resource budgets. To address this, Azar et al. [5] recently proposed an optimization strategy LambdaCrawl aiming to maximize content freshness within a given resource budget. In this paper, we demonstrate that the effectiveness of LambdaCrawl is governed in large part by how well future content change rate can be estimated. By adopting the state-of-the-art deep learning models for change rate prediction, we obtain a substantial increase of content freshness over the common LambdaCrawl implementation with change rate estimated from the past history. Moreover, we demonstrate that while LambdaCrawl is a significant advancement upon existing recrawl strategies, it can be further improved upon by a unified multi-strategy recrawl policy. To this end, we adopt the K-armed adversarial bandits algorithm that can provably optimize the overall freshness by combining multiple strategies. Empirical results over a large-scale production dataset confirm its superiority to LambdaCrawl, especially under tight resource budgets.
{"title":"Adversarial Bandits Policy for Crawling Commercial Web Content","authors":"Shuguang Han, Michael Bendersky, Przemek Gajda, Sergey Novikov, Marc Najork, Bernhard Brodowsky, Alexandrin Popescul","doi":"10.1145/3366423.3380125","DOIUrl":"https://doi.org/10.1145/3366423.3380125","url":null,"abstract":"The rapid growth of commercial web content has driven the development of shopping search services to help users find product offers. Due to the dynamic nature of commercial content, an effective recrawl policy is a key component in a shopping search service; it ensures that users have access to the up-to-date product details. Most of the existing strategies either relied on simple heuristics, or overlooked the resource budgets. To address this, Azar et al. [5] recently proposed an optimization strategy LambdaCrawl aiming to maximize content freshness within a given resource budget. In this paper, we demonstrate that the effectiveness of LambdaCrawl is governed in large part by how well future content change rate can be estimated. By adopting the state-of-the-art deep learning models for change rate prediction, we obtain a substantial increase of content freshness over the common LambdaCrawl implementation with change rate estimated from the past history. Moreover, we demonstrate that while LambdaCrawl is a significant advancement upon existing recrawl strategies, it can be further improved upon by a unified multi-strategy recrawl policy. To this end, we adopt the K-armed adversarial bandits algorithm that can provably optimize the overall freshness by combining multiple strategies. Empirical results over a large-scale production dataset confirm its superiority to LambdaCrawl, especially under tight resource budgets.","PeriodicalId":20754,"journal":{"name":"Proceedings of The Web Conference 2020","volume":"148 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2020-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89072482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper, we aim to understand the distribution of the densest subgraphs of a given graph under the density notion of average-degree. We show that the structures, the relationships and the distributions of all the densest subgraphs of a graph G can be encoded in O(L) space in an index called the ds-Index. Here L denotes the maximum output size of a densest subgraph of G. More importantly, ds-Indexcan report all the minimal densest subgraphs of G collectively in O(L) time and can enumerate all the densest subgraphs of G with an O(L) delay. Besides, the construction of ds-Indexcosts no more than finding a single densest subgraph using the state-of-the-art approach. Our empirical study shows that for web-scale graphs with one billion edges, the ds-Indexcan be constructed in several minutes on an ordinary commercial machine.
{"title":"Deconstruct Densest Subgraphs","authors":"Lijun Chang, Miao Qiao","doi":"10.1145/3366423.3380033","DOIUrl":"https://doi.org/10.1145/3366423.3380033","url":null,"abstract":"In this paper, we aim to understand the distribution of the densest subgraphs of a given graph under the density notion of average-degree. We show that the structures, the relationships and the distributions of all the densest subgraphs of a graph G can be encoded in O(L) space in an index called the ds-Index. Here L denotes the maximum output size of a densest subgraph of G. More importantly, ds-Indexcan report all the minimal densest subgraphs of G collectively in O(L) time and can enumerate all the densest subgraphs of G with an O(L) delay. Besides, the construction of ds-Indexcosts no more than finding a single densest subgraph using the state-of-the-art approach. Our empirical study shows that for web-scale graphs with one billion edges, the ds-Indexcan be constructed in several minutes on an ordinary commercial machine.","PeriodicalId":20754,"journal":{"name":"Proceedings of The Web Conference 2020","volume":"25 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2020-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79091010","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Brown Farinholt, Mohammad Rezaeirad, Damon McCoy, Kirill Levchenko
Remote Access Trojans (RATs) are a persistent class of malware that give an attacker direct, interactive access to a victim’s personal computer, allowing the attacker to steal private data, spy on the victim in real-time using the camera and microphone, and verbally harass the victim through the speaker. To date, the users and victims of this pernicious form of malware have been challenging to observe in the wild due to the unobtrusive nature of infections. In this work, we report the results of a longitudinal study of the DarkComet RAT ecosystem. Using a known method for collecting victim log databases from DarkComet controllers, we present novel techniques for tracking RAT controllers across hostname changes and improve on established techniques for filtering spurious victim records caused by scanners and sandboxed malware executions. We downloaded 6,620 DarkComet databases from 1,029 unique controllers spanning over 5 years of operation. Our analysis shows that there have been at least 57,805 victims of DarkComet over this period, with 69 new victims infected every day; many of whose keystrokes have been captured, actions recorded, and webcams monitored during this time. Our methodologies for more precisely identifying campaigns and victims could potentially be useful for improving the efficiency and efficacy of victim cleanup efforts and prioritization of law enforcement investigations.
{"title":"Dark Matter: Uncovering the DarkComet RAT Ecosystem","authors":"Brown Farinholt, Mohammad Rezaeirad, Damon McCoy, Kirill Levchenko","doi":"10.1145/3366423.3380277","DOIUrl":"https://doi.org/10.1145/3366423.3380277","url":null,"abstract":"Remote Access Trojans (RATs) are a persistent class of malware that give an attacker direct, interactive access to a victim’s personal computer, allowing the attacker to steal private data, spy on the victim in real-time using the camera and microphone, and verbally harass the victim through the speaker. To date, the users and victims of this pernicious form of malware have been challenging to observe in the wild due to the unobtrusive nature of infections. In this work, we report the results of a longitudinal study of the DarkComet RAT ecosystem. Using a known method for collecting victim log databases from DarkComet controllers, we present novel techniques for tracking RAT controllers across hostname changes and improve on established techniques for filtering spurious victim records caused by scanners and sandboxed malware executions. We downloaded 6,620 DarkComet databases from 1,029 unique controllers spanning over 5 years of operation. Our analysis shows that there have been at least 57,805 victims of DarkComet over this period, with 69 new victims infected every day; many of whose keystrokes have been captured, actions recorded, and webcams monitored during this time. Our methodologies for more precisely identifying campaigns and victims could potentially be useful for improving the efficiency and efficacy of victim cleanup efforts and prioritization of law enforcement investigations.","PeriodicalId":20754,"journal":{"name":"Proceedings of The Web Conference 2020","volume":"5 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2020-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73131862","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Guiliang Liu, Xu Li, Jiakang Wang, Mingming Sun, P. Li
To extract knowledge from general web text, it requires to build a domain-independent extractor that scales to the entire web corpus. This task is known as Open Information Extraction (OIE). This paper proposes to apply Monte-Carlo Tree Search (MCTS) to accomplish OIE. To achieve this goal, we define a Markov Decision Process for OIE and build a simulator to learn the reward signals, which provides a complete reinforcement learning framework for MCTS. Using this framework, MCTS explores candidate words (and symbols) under the guidance of a pre-trained Sequence-to-Sequence (Seq2Seq) predictor and generates abundant exploration samples during training. We apply the exploration samples to update the reward simulator and the predictor, based on which we implement another MCTS to search the optimal predictions during inference. Empirical evaluation demonstrates that the MCTS inference substantially improves the accuracy of prediction (more than 10%) and achieves a leading performance over other state-of-the-art comparison models.
{"title":"Extracting Knowledge from Web Text with Monte Carlo Tree Search","authors":"Guiliang Liu, Xu Li, Jiakang Wang, Mingming Sun, P. Li","doi":"10.1145/3366423.3380010","DOIUrl":"https://doi.org/10.1145/3366423.3380010","url":null,"abstract":"To extract knowledge from general web text, it requires to build a domain-independent extractor that scales to the entire web corpus. This task is known as Open Information Extraction (OIE). This paper proposes to apply Monte-Carlo Tree Search (MCTS) to accomplish OIE. To achieve this goal, we define a Markov Decision Process for OIE and build a simulator to learn the reward signals, which provides a complete reinforcement learning framework for MCTS. Using this framework, MCTS explores candidate words (and symbols) under the guidance of a pre-trained Sequence-to-Sequence (Seq2Seq) predictor and generates abundant exploration samples during training. We apply the exploration samples to update the reward simulator and the predictor, based on which we implement another MCTS to search the optimal predictions during inference. Empirical evaluation demonstrates that the MCTS inference substantially improves the accuracy of prediction (more than 10%) and achieves a leading performance over other state-of-the-art comparison models.","PeriodicalId":20754,"journal":{"name":"Proceedings of The Web Conference 2020","volume":"27 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2020-04-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74328538","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：