Approaches to adversarial drift

Authors:
Alex Kantchelian;Sadia Afroz;Ling Huang;Aylin Caliskan Islam;Brad Miller;Michael Carl Tschantz;Rachel Greenstadt;Anthony D. Joseph;J. D. Tygar
Affiliations:
University of California at Berkeley, Berkeley, CA, USA;University of Drexel, Philadelphia, PA, USA;Intel Labs, Berkeley, CA, USA;University of Drexel, Philadelphia, PA, USA;University of California at Berkeley, Berkeley, CA, USA;University of California at Berkeley, Berkeley, CA, USA;University of Drexel, Philadephia, PA, USA;University of California at Berkeley, Berkeley, CA, USA;University of California at Berkeley, Berkeley, CA, USA
Venue:
Proceedings of the 2013 ACM workshop on Artificial intelligence and security
Year:
2013

Citing 34
Cited 0

On the exponential value of labeled samples

Pattern Recognition Letters
Machine Learning

Machine Learning
Throttling Viruses: Restricting propagation to defeat malicious mobile code

ACSAC '02 Proceedings of the 18th Annual Computer Security Applications Conference
Adversarial classification

Proceedings of the tenth ACM SIGKDD international conference on Knowledge discovery and data mining
Polygraph: Automatically Generating Signatures for Polymorphic Worms

SP '05 Proceedings of the 2005 IEEE Symposium on Security and Privacy
Filtering spam with behavioral blacklisting

Proceedings of the 14th ACM conference on Computer and communications security
Exploiting machine learning to subvert your spam filter

LEET'08 Proceedings of the 1st Usenix Workshop on Large-Scale Exploits and Emergent Threats
Casting out Demons: Sanitizing Training Data for Anomaly Sensors

SP '08 Proceedings of the 2008 IEEE Symposium on Security and Privacy
Learning and Classification of Malware Behavior

DIMVA '08 Proceedings of the 5th international conference on Detection of Intrusions and Malware, and Vulnerability Assessment
LIBLINEAR: A Library for Large Linear Classification

The Journal of Machine Learning Research
Combining Online Classification Approaches for Changing Environments

SSPR & SPR '08 Proceedings of the 2008 Joint IAPR International Workshop on Structural, Syntactic, and Statistical Pattern Recognition
ANTIDOTE: understanding and defending against poisoning of anomaly detectors

Proceedings of the 9th ACM SIGCOMM conference on Internet measurement conference
An empirical study of malware evolution

COMSNETS'09 Proceedings of the First international conference on COMmunication Systems And NETworks
Ensemble-based classifiers

Artificial Intelligence Review
Quality management on Amazon Mechanical Turk

Proceedings of the ACM SIGKDD Workshop on Human Computation
Mining adversarial patterns via regularized loss minimization

Machine Learning
Outside the Closed World: On Using Machine Learning for Network Intrusion Detection

SP '10 Proceedings of the 2010 IEEE Symposium on Security and Privacy
Facebook immune system

Proceedings of the 4th Workshop on Social Network Systems
Design and Evaluation of a Real-Time URL Spam Filtering Service

SP '11 Proceedings of the 2011 IEEE Symposium on Security and Privacy
Automatic analysis of malware behavior using machine learning

Journal of Computer Security
Detecting adversarial advertisements in the wild

Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining
Stackelberg games for adversarial prediction problems

Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining
ZOZZLE: fast and precise in-browser JavaScript malware detection

SEC'11 Proceedings of the 20th USENIX conference on Security
Bagging classifiers for fighting poisoning attacks in adversarial classification tasks

MCS'11 Proceedings of the 10th international conference on Multiple classifier systems
Adversarial machine learning

Proceedings of the 4th ACM workshop on Security and artificial intelligence
SURF: detecting and measuring search poisoning

Proceedings of the 18th ACM conference on Computer and communications security
Paragraph: thwarting signature learning by training maliciously

RAID'06 Proceedings of the 9th international conference on Recent Advances in Intrusion Detection
Peek-a-Boo, I Still See You: Why Efficient Traffic Analysis Countermeasures Fail

SP '12 Proceedings of the 2012 IEEE Symposium on Security and Privacy
A comparison of methods for multiclass support vector machines

IEEE Transactions on Neural Networks
Robust detection of comment spam using entropy rate

Proceedings of the 5th ACM workshop on Security and artificial intelligence
Tracking concept drift in malware families

Proceedings of the 5th ACM workshop on Security and artificial intelligence
Autonomous learning for detection of JavaScript attacks: vision or reality?

Proceedings of the 5th ACM workshop on Security and artificial intelligence
Knowing your enemy: understanding and detecting malicious web advertising

Proceedings of the 2012 ACM conference on Computer and communications security
Static prediction games for adversarial learning problems

The Journal of Machine Learning Research

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this position paper, we argue that to be of practical interest, a machine-learning based security system must engage with the human operators beyond feature engineering and instance labeling to address the challenge of drift in adversarial environments. We propose that designers of such systems broaden the classification goal into an explanatory goal, which would deepen the interaction with system's operators. To provide guidance, we advocate for an approach based on maintaining one classifier for each class of unwanted activity to be filtered. We also emphasize the necessity for the system to be responsive to the operators constant curation of the training set. We show how this paradigm provides a property we call isolation and how it relates to classical causative attacks. In order to demonstrate the effects of drift on a binary classification task, we also report on two experiments using a previously unpublished malware data set where each instance is timestamped according to when it was seen.