Visible to the public Biblio

Filters: Keyword is AI Poisoning  [Clear All Filters]
2020-11-04
Apruzzese, G., Colajanni, M., Ferretti, L., Marchetti, M..  2019.  Addressing Adversarial Attacks Against Security Systems Based on Machine Learning. 2019 11th International Conference on Cyber Conflict (CyCon). 900:1—18.

Machine-learning solutions are successfully adopted in multiple contexts but the application of these techniques to the cyber security domain is complex and still immature. Among the many open issues that affect security systems based on machine learning, we concentrate on adversarial attacks that aim to affect the detection and prediction capabilities of machine-learning models. We consider realistic types of poisoning and evasion attacks targeting security solutions devoted to malware, spam and network intrusion detection. We explore the possible damages that an attacker can cause to a cyber detector and present some existing and original defensive techniques in the context of intrusion detection systems. This paper contains several performance evaluations that are based on extensive experiments using large traffic datasets. The results highlight that modern adversarial attacks are highly effective against machine-learning classifiers for cyber detection, and that existing solutions require improvements in several directions. The paper paves the way for more robust machine-learning-based techniques that can be integrated into cyber security platforms.

Khalid, F., Hanif, M. A., Rehman, S., Ahmed, R., Shafique, M..  2019.  TrISec: Training Data-Unaware Imperceptible Security Attacks on Deep Neural Networks. 2019 IEEE 25th International Symposium on On-Line Testing and Robust System Design (IOLTS). :188—193.

Most of the data manipulation attacks on deep neural networks (DNNs) during the training stage introduce a perceptible noise that can be catered by preprocessing during inference, or can be identified during the validation phase. There-fore, data poisoning attacks during inference (e.g., adversarial attacks) are becoming more popular. However, many of them do not consider the imperceptibility factor in their optimization algorithms, and can be detected by correlation and structural similarity analysis, or noticeable (e.g., by humans) in multi-level security system. Moreover, majority of the inference attack rely on some knowledge about the training dataset. In this paper, we propose a novel methodology which automatically generates imperceptible attack images by using the back-propagation algorithm on pre-trained DNNs, without requiring any information about the training dataset (i.e., completely training data-unaware). We present a case study on traffic sign detection using the VGGNet trained on the German Traffic Sign Recognition Benchmarks dataset in an autonomous driving use case. Our results demonstrate that the generated attack images successfully perform misclassification while remaining imperceptible in both “subjective” and “objective” quality tests.

Chacon, H., Silva, S., Rad, P..  2019.  Deep Learning Poison Data Attack Detection. 2019 IEEE 31st International Conference on Tools with Artificial Intelligence (ICTAI). :971—978.

Deep neural networks are widely used in many walks of life. Techniques such as transfer learning enable neural networks pre-trained on certain tasks to be retrained for a new duty, often with much less data. Users have access to both pre-trained model parameters and model definitions along with testing data but have either limited access to training data or just a subset of it. This is risky for system-critical applications, where adversarial information can be maliciously included during the training phase to attack the system. Determining the existence and level of attack in a model is challenging. In this paper, we present evidence on how adversarially attacking training data increases the boundary of model parameters using as an example of a CNN model and the MNIST data set as a test. This expansion is due to new characteristics of the poisonous data that are added to the training data. Approaching the problem from the feature space learned by the network provides a relation between them and the possible parameters taken by the model on the training phase. An algorithm is proposed to determine if a given network was attacked in the training by comparing the boundaries of parameters distribution on intermediate layers of the model estimated by using the Maximum Entropy Principle and the Variational inference approach.

Zhang, J., Chen, J., Wu, D., Chen, B., Yu, S..  2019.  Poisoning Attack in Federated Learning using Generative Adversarial Nets. 2019 18th IEEE International Conference On Trust, Security And Privacy In Computing And Communications/13th IEEE International Conference On Big Data Science And Engineering (TrustCom/BigDataSE). :374—380.

Federated learning is a novel distributed learning framework, where the deep learning model is trained in a collaborative manner among thousands of participants. The shares between server and participants are only model parameters, which prevent the server from direct access to the private training data. However, we notice that the federated learning architecture is vulnerable to an active attack from insider participants, called poisoning attack, where the attacker can act as a benign participant in federated learning to upload the poisoned update to the server so that he can easily affect the performance of the global model. In this work, we study and evaluate a poisoning attack in federated learning system based on generative adversarial nets (GAN). That is, an attacker first acts as a benign participant and stealthily trains a GAN to mimic prototypical samples of the other participants' training set which does not belong to the attacker. Then these generated samples will be fully controlled by the attacker to generate the poisoning updates, and the global model will be compromised by the attacker with uploading the scaled poisoning updates to the server. In our evaluation, we show that the attacker in our construction can successfully generate samples of other benign participants using GAN and the global model performs more than 80% accuracy on both poisoning tasks and main tasks.

Liang, Y., He, D., Chen, D..  2019.  Poisoning Attack on Load Forecasting. 2019 IEEE Innovative Smart Grid Technologies - Asia (ISGT Asia). :1230—1235.

Short-term load forecasting systems for power grids have demonstrated high accuracy and have been widely employed for commercial use. However, classic load forecasting systems, which are based on statistical methods, are subject to vulnerability from training data poisoning. In this paper, we demonstrate a data poisoning strategy that effectively corrupts the forecasting model even in the presence of outlier detection. To the best of our knowledge, poisoning attack on short-term load forecasting with outlier detection has not been studied in previous works. Our method applies to several forecasting models, including the most widely-adapted and best-performing ones, such as multiple linear regression (MLR) and neural network (NN) models. Starting with the MLR model, we develop a novel closed-form solution to quickly estimate the new MLR model after a round of data poisoning without retraining. We then employ line search and simulated annealing to find the poisoning attack solution. Furthermore, we use the MLR attacking solution to generate a numerical solution for other models, such as NN. The effectiveness of our algorithm has been tested on the Global Energy Forecasting Competition (GEFCom2012) data set with the presence of outlier detection.

Kim, Y., Ahn, S., Thang, N. C., Choi, D., Park, M..  2019.  ARP Poisoning Attack Detection Based on ARP Update State in Software-Defined Networks. 2019 International Conference on Information Networking (ICOIN). :366—371.

Recently, the novel networking technology Software-Defined Networking(SDN) and Service Function Chaining(SFC) are rapidly growing, and security issues are also emerging for SDN and SFC. However, the research about security and safety on a novel networking environment is still unsatisfactory, and the vulnerabilities have been revealed continuously. Among these security issues, this paper addresses the ARP Poisoning attack to exploit SFC vulnerability, and proposes a method to defend the attack. The proposed method recognizes the repetitive ARP reply which is a feature of ARP Poisoning attack, and detects ARP Poisoning attack. The proposed method overcomes the limitations of the existing detection methods. The proposed method also detects the presence of an attack more accurately.

Jin, Y., Tomoishi, M., Matsuura, S..  2019.  A Detection Method Against DNS Cache Poisoning Attacks Using Machine Learning Techniques: Work in Progress. 2019 IEEE 18th International Symposium on Network Computing and Applications (NCA). :1—3.

DNS based domain name resolution has been known as one of the most fundamental Internet services. In the meanwhile, DNS cache poisoning attacks also have become a critical threat in the cyber world. In addition to Kaminsky attacks, the falsified data from the compromised authoritative DNS servers also have become the threats nowadays. Several solutions have been proposed in order to prevent DNS cache poisoning attacks in the literature for the former case such as DNSSEC (DNS Security Extensions), however no effective solutions have been proposed for the later case. Moreover, due to the performance issue and significant workload increase on DNS cache servers, DNSSEC has not been deployed widely yet. In this work, we propose an advanced detection method against DNS cache poisoning attacks using machine learning techniques. In the proposed method, in addition to the basic 5-tuple information of a DNS packet, we intend to add a lot of special features extracted based on the standard DNS protocols as well as the heuristic aspects such as “time related features”, “GeoIP related features” and “trigger of cached DNS data”, etc., in order to identify the DNS response packets used for cache poisoning attacks especially those from compromised authoritative DNS servers. In this paper, as a work in progress, we describe the basic idea and concept of our proposed method as well as the intended network topology of the experimental environment while the prototype implementation, training data preparation and model creation as well as the evaluations will belong to the future work.

Khurana, N., Mittal, S., Piplai, A., Joshi, A..  2019.  Preventing Poisoning Attacks On AI Based Threat Intelligence Systems. 2019 IEEE 29th International Workshop on Machine Learning for Signal Processing (MLSP). :1—6.

As AI systems become more ubiquitous, securing them becomes an emerging challenge. Over the years, with the surge in online social media use and the data available for analysis, AI systems have been built to extract, represent and use this information. The credibility of this information extracted from open sources, however, can often be questionable. Malicious or incorrect information can cause a loss of money, reputation, and resources; and in certain situations, pose a threat to human life. In this paper, we use an ensembled semi-supervised approach to determine the credibility of Reddit posts by estimating their reputation score to ensure the validity of information ingested by AI systems. We demonstrate our approach in the cybersecurity domain, where security analysts utilize these systems to determine possible threats by analyzing the data scattered on social media websites, forums, blogs, etc.

Shen, J., Zhu, X., Ma, D..  2019.  TensorClog: An Imperceptible Poisoning Attack on Deep Neural Network Applications. IEEE Access. 7:41498—41506.

Internet application providers now have more incentive than ever to collect user data, which greatly increases the risk of user privacy violations due to the emerging of deep neural networks. In this paper, we propose TensorClog-a poisoning attack technique that is designed for privacy protection against deep neural networks. TensorClog has three properties with each of them serving a privacy protection purpose: 1) training on TensorClog poisoned data results in lower inference accuracy, reducing the incentive of abusive data collection; 2) training on TensorClog poisoned data converges to a larger loss, which prevents the neural network from learning the privacy; and 3) TensorClog regularizes the perturbation to remain a high structure similarity, so that the poisoning does not affect the actual content in the data. Applying our TensorClog poisoning technique to CIFAR-10 dataset results in an increase in both converged training loss and test error by 300% and 272%, respectively. It manages to maintain data's human perception with a high SSIM index of 0.9905. More experiments including different limited information attack scenarios and a real-world application transferred from pre-trained ImageNet models are presented to further evaluate TensorClog's effectiveness in more complex situations.

2020-08-10
Kwon, Hyun, Yoon, Hyunsoo, Park, Ki-Woong.  2019.  Selective Poisoning Attack on Deep Neural Network to Induce Fine-Grained Recognition Error. 2019 IEEE Second International Conference on Artificial Intelligence and Knowledge Engineering (AIKE). :136–139.

Deep neural networks (DNNs) provide good performance for image recognition, speech recognition, and pattern recognition. However, a poisoning attack is a serious threat to DNN's security. The poisoning attack is a method to reduce the accuracy of DNN by adding malicious training data during DNN training process. In some situations such as a military, it may be necessary to drop only a chosen class of accuracy in the model. For example, if an attacker does not allow only nuclear facilities to be selectively recognized, it may be necessary to intentionally prevent UAV from correctly recognizing nuclear-related facilities. In this paper, we propose a selective poisoning attack that reduces the accuracy of only chosen class in the model. The proposed method reduces the accuracy of a chosen class in the model by training malicious training data corresponding to a chosen class, while maintaining the accuracy of the remaining classes. For experiment, we used tensorflow as a machine learning library and MNIST and CIFAR10 as datasets. Experimental results show that the proposed method can reduce the accuracy of the chosen class to 43.2% and 55.3% in MNIST and CIFAR10, while maintaining the accuracy of the remaining classes.

2018-07-06
Kloft, Marius, Laskov, Pavel.  2012.  Security Analysis of Online Centroid Anomaly Detection. J. Mach. Learn. Res.. 13:3681–3724.

Security issues are crucial in a number of machine learning applications, especially in scenarios dealing with human activity rather than natural phenomena (e.g., information ranking, spam detection, malware detection, etc.). In such cases, learning algorithms may have to cope with manipulated data aimed at hampering decision making. Although some previous work addressed the issue of handling malicious data in the context of supervised learning, very little is known about the behavior of anomaly detection methods in such scenarios. In this contribution, we analyze the performance of a particular method–online centroid anomaly detection–in the presence of adversarial noise. Our analysis addresses the following security-related issues: formalization of learning and attack processes, derivation of an optimal attack, and analysis of attack efficiency and limitations. We derive bounds on the effectiveness of a poisoning attack against centroid anomaly detection under different conditions: attacker's full or limited control over the traffic and bounded false positive rate. Our bounds show that whereas a poisoning attack can be effectively staged in the unconstrained case, it can be made arbitrarily difficult (a strict upper bound on the attacker's gain) if external constraints are properly used. Our experimental evaluation, carried out on real traces of HTTP and exploit traffic, confirms the tightness of our theoretical bounds and the practicality of our protection mechanisms.

Baracaldo, Nathalie, Chen, Bryant, Ludwig, Heiko, Safavi, Jaehoon Amir.  2017.  Mitigating Poisoning Attacks on Machine Learning Models: A Data Provenance Based Approach. Proceedings of the 10th ACM Workshop on Artificial Intelligence and Security. :103–110.
The use of machine learning models has become ubiquitous. Their predictions are used to make decisions about healthcare, security, investments and many other critical applications. Given this pervasiveness, it is not surprising that adversaries have an incentive to manipulate machine learning models to their advantage. One way of manipulating a model is through a poisoning or causative attack in which the adversary feeds carefully crafted poisonous data points into the training set. Taking advantage of recently developed tamper-free provenance frameworks, we present a methodology that uses contextual information about the origin and transformation of data points in the training set to identify poisonous data, thereby enabling online and regularly re-trained machine learning applications to consume data sources in potentially adversarial environments. To the best of our knowledge, this is the first approach to incorporate provenance information as part of a filtering algorithm to detect causative attacks. We present two variations of the methodology - one tailored to partially trusted data sets and the other to fully untrusted data sets. Finally, we evaluate our methodology against existing methods to detect poison data and show an improvement in the detection rate.
Liu, Chang, Li, Bo, Vorobeychik, Yevgeniy, Oprea, Alina.  2017.  Robust Linear Regression Against Training Data Poisoning. Proceedings of the 10th ACM Workshop on Artificial Intelligence and Security. :91–102.
The effectiveness of supervised learning techniques has made them ubiquitous in research and practice. In high-dimensional settings, supervised learning commonly relies on dimensionality reduction to improve performance and identify the most important factors in predicting outcomes. However, the economic importance of learning has made it a natural target for adversarial manipulation of training data, which we term poisoning attacks. Prior approaches to dealing with robust supervised learning rely on strong assumptions about the nature of the feature matrix, such as feature independence and sub-Gaussian noise with low variance. We propose an integrated method for robust regression that relaxes these assumptions, assuming only that the feature matrix can be well approximated by a low-rank matrix. Our techniques integrate improved robust low-rank matrix approximation and robust principle component regression, and yield strong performance guarantees. Moreover, we experimentally show that our methods significantly outperform state of the art both in running time and prediction error.
Biggio, Battista, Rieck, Konrad, Ariu, Davide, Wressnegger, Christian, Corona, Igino, Giacinto, Giorgio, Roli, Fabio.  2014.  Poisoning Behavioral Malware Clustering. Proceedings of the 2014 Workshop on Artificial Intelligent and Security Workshop. :27–36.
Clustering algorithms have become a popular tool in computer security to analyze the behavior of malware variants, identify novel malware families, and generate signatures for antivirus systems. However, the suitability of clustering algorithms for security-sensitive settings has been recently questioned by showing that they can be significantly compromised if an attacker can exercise some control over the input data. In this paper, we revisit this problem by focusing on behavioral malware clustering approaches, and investigate whether and to what extent an attacker may be able to subvert these approaches through a careful injection of samples with poisoning behavior. To this end, we present a case study on Malheur, an open-source tool for behavioral malware clustering. Our experiments not only demonstrate that this tool is vulnerable to poisoning attacks, but also that it can be significantly compromised even if the attacker can only inject a very small percentage of attacks into the input data. As a remedy, we discuss possible countermeasures and highlight the need for more secure clustering algorithms.
Liu, T., Wen, W., Jin, Y..  2018.  SIN2: Stealth infection on neural network \#x2014; A low-cost agile neural Trojan attack methodology. 2018 IEEE International Symposium on Hardware Oriented Security and Trust (HOST). :227–230.

Deep Neural Network (DNN) has recently become the “de facto” technique to drive the artificial intelligence (AI) industry. However, there also emerges many security issues as the DNN based intelligent systems are being increasingly prevalent. Existing DNN security studies, such as adversarial attacks and poisoning attacks, are usually narrowly conducted at the software algorithm level, with the misclassification as their primary goal. The more realistic system-level attacks introduced by the emerging intelligent service supply chain, e.g. the third-party cloud based machine learning as a service (MLaaS) along with the portable DNN computing engine, have never been discussed. In this work, we propose a low-cost modular methodology-Stealth Infection on Neural Network, namely “SIN2”, to demonstrate the novel and practical intelligent supply chain triggered neural Trojan attacks. Our “SIN2” well leverages the attacking opportunities built upon the static neural network model and the underlying dynamic runtime system of neural computing framework through a bunch of neural Trojaning techniques. We implement a variety of neural Trojan attacks in Linux sandbox by following proposed “SIN2”. Experimental results show that our modular design can rapidly produce and trigger various Trojan attacks that can easily evade the existing defenses.

Du, Xiaojiang.  2004.  Using k-nearest neighbor method to identify poison message failure. IEEE Global Telecommunications Conference, 2004. GLOBECOM '04. 4:2113–2117Vol.4.

Poison message failure is a mechanism that has been responsible for large scale failures in both telecommunications and IP networks. The poison message failure can propagate in the network and cause an unstable network. We apply a machine learning, data mining technique in the network fault management area. We use the k-nearest neighbor method to identity the poison message failure. We also propose a "probabilistic" k-nearest neighbor method which outputs a probability distribution about the poison message. Through extensive simulations, we show that the k-nearest neighbor method is very effective in identifying the responsible message type.

Mozaffari-Kermani, M., Sur-Kolay, S., Raghunathan, A., Jha, N. K..  2015.  Systematic Poisoning Attacks on and Defenses for Machine Learning in Healthcare. IEEE Journal of Biomedical and Health Informatics. 19:1893–1905.

Machine learning is being used in a wide range of application domains to discover patterns in large datasets. Increasingly, the results of machine learning drive critical decisions in applications related to healthcare and biomedicine. Such health-related applications are often sensitive, and thus, any security breach would be catastrophic. Naturally, the integrity of the results computed by machine learning is of great importance. Recent research has shown that some machine-learning algorithms can be compromised by augmenting their training datasets with malicious data, leading to a new class of attacks called poisoning attacks. Hindrance of a diagnosis may have life-threatening consequences and could cause distrust. On the other hand, not only may a false diagnosis prompt users to distrust the machine-learning algorithm and even abandon the entire system but also such a false positive classification may cause patient distress. In this paper, we present a systematic, algorithm-independent approach for mounting poisoning attacks across a wide range of machine-learning algorithms and healthcare datasets. The proposed attack procedure generates input data, which, when added to the training set, can either cause the results of machine learning to have targeted errors (e.g., increase the likelihood of classification into a specific class), or simply introduce arbitrary errors (incorrect classification). These attacks may be applied to both fixed and evolving datasets. They can be applied even when only statistics of the training dataset are available or, in some cases, even without access to the training dataset, although at a lower efficacy. We establish the effectiveness of the proposed attacks using a suite of six machine-learning algorithms and five healthcare datasets. Finally, we present countermeasures against the proposed generic attacks that are based on tracking and detecting deviations in various accuracy metrics, and benchmark their effectiveness.

Zhang, F., Chan, P. P. K., Tang, T. Q..  2015.  L-GEM based robust learning against poisoning attack. 2015 International Conference on Wavelet Analysis and Pattern Recognition (ICWAPR). :175–178.

Poisoning attack in which an adversary misleads the learning process by manipulating its training set significantly affect the performance of classifiers in security applications. This paper proposed a robust learning method which reduces the influences of attack samples on learning. The sensitivity, defined as the fluctuation of the output with small perturbation of the input, in Localized Generalization Error Model (L-GEM) is measured for each training sample. The classifier's output on attack samples may be sensitive and inaccurate since these samples are different from other untainted samples. An import score is assigned to each sample according to its localized generalization error bound. The classifier is trained using a new training set obtained by resampling the samples according to their importance scores. RBFNN is applied as the classifier in experimental evaluation. The proposed model outperforms than the traditional one under the well-known label flip poisoning attacks including nearest-first and farthest-first flips attack.

Sun, R., Yuan, X., Lee, A., Bishop, M., Porter, D. E., Li, X., Gregio, A., Oliveira, D..  2017.  The dose makes the poison \#x2014; Leveraging uncertainty for effective malware detection. 2017 IEEE Conference on Dependable and Secure Computing. :123–130.

Malware has become sophisticated and organizations don't have a Plan B when standard lines of defense fail. These failures have devastating consequences for organizations, such as sensitive information being exfiltrated. A promising avenue for improving the effectiveness of behavioral-based malware detectors is to combine fast (usually not highly accurate) traditional machine learning (ML) detectors with high-accuracy, but time-consuming, deep learning (DL) models. The main idea is to place software receiving borderline classifications by traditional ML methods in an environment where uncertainty is added, while software is analyzed by time-consuming DL models. The goal of uncertainty is to rate-limit actions of potential malware during deep analysis. In this paper, we describe Chameleon, a Linux-based framework that implements this uncertain environment. Chameleon offers two environments for its OS processes: standard - for software identified as benign by traditional ML detectors - and uncertain - for software that received borderline classifications analyzed by ML methods. The uncertain environment will bring obstacles to software execution through random perturbations applied probabilistically on selected system calls. We evaluated Chameleon with 113 applications from common benchmarks and 100 malware samples for Linux. Our results show that at threshold 10%, intrusive and non-intrusive strategies caused approximately 65% of malware to fail accomplishing their tasks, while approximately 30% of the analyzed benign software to meet with various levels of disruption (crashed or hampered). We also found that I/O-bound software was three times more affected by uncertainty than CPU-bound software.

Zhang, R., Zhu, Q..  2017.  A game-theoretic defense against data poisoning attacks in distributed support vector machines. 2017 IEEE 56th Annual Conference on Decision and Control (CDC). :4582–4587.

With a large number of sensors and control units in networked systems, distributed support vector machines (DSVMs) play a fundamental role in scalable and efficient multi-sensor classification and prediction tasks. However, DSVMs are vulnerable to adversaries who can modify and generate data to deceive the system to misclassification and misprediction. This work aims to design defense strategies for DSVM learner against a potential adversary. We use a game-theoretic framework to capture the conflicting interests between the DSVM learner and the attacker. The Nash equilibrium of the game allows predicting the outcome of learning algorithms in adversarial environments, and enhancing the resilience of the machine learning through dynamic distributed algorithms. We develop a secure and resilient DSVM algorithm with rejection method, and show its resiliency against adversary with numerical experiments.

Lampesberger, H..  2016.  An Incremental Learner for Language-Based Anomaly Detection in XML. 2016 IEEE Security and Privacy Workshops (SPW). :156–170.

The Extensible Markup Language (XML) is a complex language, and consequently, XML-based protocols are susceptible to entire classes of implicit and explicit security problems. Message formats in XML-based protocols are usually specified in XML Schema, and as a first-line defense, schema validation should reject malformed input. However, extension points in most protocol specifications break validation. Extension points are wildcards and considered best practice for loose composition, but they also enable an attacker to add unchecked content in a document, e.g., for a signature wrapping attack. This paper introduces datatyped XML visibly pushdown automata (dXVPAs) as language representation for mixed-content XML and presents an incremental learner that infers a dXVPA from example documents. The learner generalizes XML types and datatypes in terms of automaton states and transitions, and an inferred dXVPA converges to a good-enough approximation of the true language. The automaton is free from extension points and capable of stream validation, e.g., as an anomaly detector for XML-based protocols. For dealing with adversarial training data, two scenarios of poisoning are considered: a poisoning attack is either uncovered at a later time or remains hidden. Unlearning can therefore remove an identified poisoning attack from a dXVPA, and sanitization trims low-frequent states and transitions to get rid of hidden attacks. All algorithms have been evaluated in four scenarios, including a web service implemented in Apache Axis2 and Apache Rampart, where attacks have been simulated. In all scenarios, the learned automaton had zero false positives and outperformed traditional schema validation.

2018-04-11
Muñoz-González, Luis, Biggio, Battista, Demontis, Ambra, Paudice, Andrea, Wongrassamee, Vasin, Lupu, Emil C., Roli, Fabio.  2017.  Towards Poisoning of Deep Learning Algorithms with Back-Gradient Optimization. Proceedings of the 10th ACM Workshop on Artificial Intelligence and Security. :27–38.

A number of online services nowadays rely upon machine learning to extract valuable information from data collected in the wild. This exposes learning algorithms to the threat of data poisoning, i.e., a coordinate attack in which a fraction of the training data is controlled by the attacker and manipulated to subvert the learning process. To date, these attacks have been devised only against a limited class of binary learning algorithms, due to the inherent complexity of the gradient-based procedure used to optimize the poisoning points (a.k.a. adversarial training examples). In this work, we first extend the definition of poisoning attacks to multiclass problems. We then propose a novel poisoning algorithm based on the idea of back-gradient optimization, i.e., to compute the gradient of interest through automatic differentiation, while also reversing the learning procedure to drastically reduce the attack complexity. Compared to current poisoning strategies, our approach is able to target a wider class of learning algorithms, trained with gradient-based procedures, including neural networks and deep learning architectures. We empirically evaluate its effectiveness on several application examples, including spam filtering, malware detection, and handwritten digit recognition. We finally show that, similarly to adversarial test examples, adversarial training examples can also be transferred across different learning algorithms.