arXiv subjects

Prateek Mittal

Publications and source records attributed to Prateek Mittal.

At least 91 records · Page 5Linked to original sources

arXiv2021

PatchGuard++: Efficient Provable Attack Detection against Adversarial Patches

An adversarial patch can arbitrarily manipulate image pixels within a restricted region to induce model misclassification. The threat of this localized attack has gained significant attention because the adversary can mount a physically-realizable attack by attaching patches to the victim object. Recent provably robust defenses generally follow the PatchGuard framework by using CNNs with small receptive fields and secure feature aggregation for robust model predictions. In this paper, we extend PatchGuard to PatchGuard++ for provably detecting the adversarial patch attack to boost both provable robust accuracy and clean accuracy. In PatchGuard++, we first use a CNN with small receptive fields for feature extraction so that the number of features corrupted by the adversarial patch is bounded. Next, we apply masks in the feature space and evaluate predictions on all possible masked feature maps. Finally, we extract a pattern from all masked predictions to catch the adversarial patch attack. We evaluate PatchGuard++ on ImageNette (a 10-class subset of ImageNet), ImageNet, and CIFAR-10 and demonstrate that PatchGuard++ significantly improves the provable robustness and clean performance.

cs.CV↗

arXiv2021

PatchGuard: A Provably Robust Defense against Adversarial Patches via Small Receptive Fields and Masking

Localized adversarial patches aim to induce misclassification in machine learning models by arbitrarily modifying pixels within a restricted region of an image. Such attacks can be realized in the physical world by attaching the adversarial patch to the object to be misclassified, and defending against such attacks is an unsolved/open problem. In this paper, we propose a general defense framework called PatchGuard that can achieve high provable robustness while maintaining high clean accuracy against localized adversarial patches. The cornerstone of PatchGuard involves the use of CNNs with small receptive fields to impose a bound on the number of features corrupted by an adversarial patch. Given a bounded number of corrupted features, the problem of designing an adversarial patch defense reduces to that of designing a secure feature aggregation mechanism. Towards this end, we present our robust masking defense that robustly detects and masks corrupted features to recover the correct prediction. Notably, we can prove the robustness of our defense against any adversary within our threat model. Our extensive evaluation on ImageNet, ImageNette (a 10-class subset of ImageNet), and CIFAR-10 datasets demonstrates that our defense achieves state-of-the-art performance in terms of both provable robust accuracy and clean accuracy.

cs.CV↗

arXiv2021

SSD: A Unified Framework for Self-Supervised Outlier Detection

We ask the following question: what training information is required to design an effective outlier/out-of-distribution (OOD) detector, i.e., detecting samples that lie far away from the training distribution? Since unlabeled data is easily accessible for many applications, the most compelling approach is to develop detectors based on only unlabeled in-distribution data. However, we observe that most existing detectors based on unlabeled data perform poorly, often equivalent to a random prediction. In contrast, existing state-of-the-art OOD detectors achieve impressive performance but require access to fine-grained data labels for supervised training. We propose SSD, an outlier detector based on only unlabeled in-distribution data. We use self-supervised representation learning followed by a Mahalanobis distance based detection in the feature space. We demonstrate that SSD outperforms most existing detectors based on unlabeled data by a large margin. Additionally, SSD even achieves performance on par, and sometimes even better, with supervised training based detectors. Finally, we expand our detection framework with two key extensions. First, we formulate few-shot OOD detection, in which the detector has access to only one to five samples from each class of the targeted OOD dataset. Second, we extend our framework to incorporate training data labels, if available. We find that our novel detection framework based on SSD displays enhanced performance with these extensions, and achieves state-of-the-art performance. Our code is publicly available at https://github.com/inspire-group/SSD.

cs.CV↗

arXiv2021

Advances and Open Problems in Federated Learning

Federated learning (FL) is a machine learning setting where many clients (e.g. mobile devices or whole organizations) collaboratively train a model under the orchestration of a central server (e.g. service provider), while keeping the training data decentralized. FL embodies the principles of focused data collection and minimization, and can mitigate many of the systemic privacy risks and costs resulting from traditional, centralized machine learning and data science approaches. Motivated by the explosive growth in FL research, this paper discusses recent advances and presents an extensive collection of open problems and challenges.

cs.LG↗

arXiv2021

Enabling Efficient Cyber Threat Hunting With Cyber Threat Intelligence

Log-based cyber threat hunting has emerged as an important solution to counter sophisticated attacks. However, existing approaches require non-trivial efforts of manual query construction and have overlooked the rich external threat knowledge provided by open-source Cyber Threat Intelligence (OSCTI). To bridge the gap, we propose ThreatRaptor, a system that facilitates threat hunting in computer systems using OSCTI. Built upon system auditing frameworks, ThreatRaptor provides (1) an unsupervised, light-weight, and accurate NLP pipeline that extracts structured threat behaviors from unstructured OSCTI text, (2) a concise and expressive domain-specific query language, TBQL, to hunt for malicious system activities, (3) a query synthesis mechanism that automatically synthesizes a TBQL query for hunting, and (4) an efficient query execution engine to search the big audit logging data. Evaluations on a broad set of attack cases demonstrate the accuracy and efficiency of ThreatRaptor in practical threat hunting.

cs.CR↗

arXiv2021

A System for Efficiently Hunting for Cyber Threats in Computer Systems Using Threat Intelligence

Log-based cyber threat hunting has emerged as an important solution to counter sophisticated cyber attacks. However, existing approaches require non-trivial efforts of manual query construction and have overlooked the rich external knowledge about threat behaviors provided by open-source Cyber Threat Intelligence (OSCTI). To bridge the gap, we build ThreatRaptor, a system that facilitates cyber threat hunting in computer systems using OSCTI. Built upon mature system auditing frameworks, ThreatRaptor provides (1) an unsupervised, light-weight, and accurate NLP pipeline that extracts structured threat behaviors from unstructured OSCTI text, (2) a concise and expressive domain-specific query language, TBQL, to hunt for malicious system activities, (3) a query synthesis mechanism that automatically synthesizes a TBQL query from the extracted threat behaviors, and (4) an efficient query execution engine to search the big system audit logging data.

cs.CR↗

arXiv2020

Systematic Evaluation of Privacy Risks of Machine Learning Models

Machine learning models are prone to memorizing sensitive data, making them vulnerable to membership inference attacks in which an adversary aims to guess if an input sample was used to train the model. In this paper, we show that prior work on membership inference attacks may severely underestimate the privacy risks by relying solely on training custom neural network classifiers to perform attacks and focusing only on the aggregate results over data samples, such as the attack accuracy. To overcome these limitations, we first propose to benchmark membership inference privacy risks by improving existing non-neural network based inference attacks and proposing a new inference attack method based on a modification of prediction entropy. We also propose benchmarks for defense mechanisms by accounting for adaptive adversaries with knowledge of the defense and also accounting for the trade-off between model accuracy and privacy risks. Using our benchmark attacks, we demonstrate that existing defense approaches are not as effective as previously reported. Next, we introduce a new approach for fine-grained privacy analysis by formulating and deriving a new metric called the privacy risk score. Our privacy risk score metric measures an individual sample's likelihood of being a training member, which allows an adversary to identify samples with high privacy risks and perform attacks with high confidence. We experimentally validate the effectiveness of the privacy risk score and demonstrate that the distribution of privacy risk score across individual samples is heterogeneous. Finally, we perform an in-depth investigation for understanding why certain samples have high privacy risks, including correlations with model sensitivity, generalization error, and feature embeddings. Our work emphasizes the importance of a systematic and rigorous evaluation of privacy risks of machine learning models.

cs.CR↗

arXiv2020

Towards Probabilistic Verification of Machine Unlearning

The right to be forgotten, also known as the right to erasure, is the right of individuals to have their data erased from an entity storing it. The status of this long held notion was legally solidified recently by the General Data Protection Regulation (GDPR) in the European Union. Consequently, there is a need for mechanisms whereby users can verify if service providers comply with their deletion requests. In this work, we take the first step in proposing a formal framework to study the design of such verification mechanisms for data deletion requests -- also known as machine unlearning -- in the context of systems that provide machine learning as a service (MLaaS). Our framework allows the rigorous quantification of any verification mechanism based on standard hypothesis testing. Furthermore, we propose a novel backdoor-based verification mechanism and demonstrate its effectiveness in certifying data deletion with high confidence, thus providing a basis for quantitatively inferring machine unlearning. We evaluate our approach over a range of network architectures such as multi-layer perceptrons (MLP), convolutional neural networks (CNN), residual networks (ResNet), and long short-term memory (LSTM), as well as over 5 different datasets. We demonstrate that our approach has minimal effect on the ML service's accuracy but provides high confidence verification of unlearning. Our proposed mechanism works even if only a handful of users employ our system to ascertain compliance with data deletion requests. In particular, with just 5% of users participating, modifying half their data with a backdoor, and with merely 30 test queries, our verification mechanism has both false positive and false negative ratios below $10^{-3}$. We also show the effectiveness of our approach by testing it against an adaptive adversary that uses a state-of-the-art backdoor defense method.

cs.CR↗

arXiv2020

HYDRA: Pruning Adversarially Robust Neural Networks

In safety-critical but computationally resource-constrained applications, deep learning faces two key challenges: lack of robustness against adversarial attacks and large neural network size (often millions of parameters). While the research community has extensively explored the use of robust training and network pruning independently to address one of these challenges, only a few recent works have studied them jointly. However, these works inherit a heuristic pruning strategy that was developed for benign training, which performs poorly when integrated with robust training techniques, including adversarial training and verifiable robust training. To overcome this challenge, we propose to make pruning techniques aware of the robust training objective and let the training objective guide the search for which connections to prune. We realize this insight by formulating the pruning objective as an empirical risk minimization problem which is solved efficiently using SGD. We demonstrate that our approach, titled HYDRA, achieves compressed networks with state-of-the-art benign and robust accuracy, simultaneously. We demonstrate the success of our approach across CIFAR-10, SVHN, and ImageNet dataset with four robust training techniques: iterative adversarial training, randomized smoothing, MixTrain, and CROWN-IBP. We also demonstrate the existence of highly robust sub-networks within non-robust networks. Our code and compressed networks are publicly available at \url{https://github.com/inspire-group/compactness-robustness}.

cs.CV↗

arXiv2020

Evaluating Snowflake as an Indistinguishable Censorship Circumvention Tool

Tor is the most well-known tool for circumventing censorship. Unfortunately, Tor traffic has been shown to be detectable using deep-packet inspection. WebRTC is a popular web frame-work that enables browser-to-browser connections. Snowflake is a novel pluggable transport that leverages WebRTC to connect Tor clients to the Tor network. In theory, Snowflake was created to be indistinguishable from other WebRTC services. In this paper, we evaluate the indistinguishability of Snowflake. We collect over 6,500 DTLS handshakes from Snowflake, Facebook Messenger, Google Hangouts, and Discord WebRTC connections and show that Snowflake is identifiable among these applications with 100% accuracy. We show that several features, including the extensions offered and the number of packets in the handshake, distinguish Snowflake among these services. Finally, we suggest recommendations for improving identification resistance in Snowflake. We have made the dataset publicly available.

cs.CR↗

arXiv2020

FALCON: Honest-Majority Maliciously Secure Framework for Private Deep Learning

We propose Falcon, an end-to-end 3-party protocol for efficient private training and inference of large machine learning models. Falcon presents four main advantages - (i) It is highly expressive with support for high capacity networks such as VGG16 (ii) it supports batch normalization which is important for training complex networks such as AlexNet (iii) Falcon guarantees security with abort against malicious adversaries, assuming an honest majority (iv) Lastly, Falcon presents new theoretical insights for protocol design that make it highly efficient and allow it to outperform existing secure deep learning solutions. Compared to prior art for private inference, we are about 8x faster than SecureNN (PETS'19) on average and comparable to ABY3 (CCS'18). We are about 16-200x more communication efficient than either of these. For private training, we are about 6x faster than SecureNN, 4.4x faster than ABY3 and about 2-60x more communication efficient. Our experiments in the WAN setting show that over large networks and datasets, compute operations dominate the overall latency of MPC, as opposed to the communication.

cs.CR↗

arXiv2020

Securing Internet Applications from Routing Attacks

Attacks on Internet routing are typically viewed through the lens of availability and confidentiality, assuming an adversary that either discards traffic or performs eavesdropping. Yet, a strategic adversary can use routing attacks to compromise the security of critical Internet applications like Tor, certificate authorities, and the bitcoin network. In this paper, we survey such application-specific routing attacks and argue that both application-layer and network-layer defenses are essential and urgently needed. While application-layer defenses are easier to deploy in the short term, we hope that our work serves to provide much needed momentum for the deployment of network-layer defenses.

cs.NI↗

arXiv2020

A Critical Evaluation of Open-World Machine Learning

Open-world machine learning (ML) combines closed-world models trained on in-distribution data with out-of-distribution (OOD) detectors, which aim to detect and reject OOD inputs. Previous works on open-world ML systems usually fail to test their reliability under diverse, and possibly adversarial conditions. Therefore, in this paper, we seek to understand how resilient are state-of-the-art open-world ML systems to changes in system components? With our evaluation across 6 OOD detectors, we find that the choice of in-distribution data, model architecture and OOD data have a strong impact on OOD detection performance, inducing false positive rates in excess of $70\%$. We further show that OOD inputs with 22 unintentional corruptions or adversarial perturbations render open-world ML systems unusable with false positive rates of up to $100\%$. To increase the resilience of open-world ML, we combine robust classifiers with OOD detection techniques and uncover a new trade-off between OOD detection and robustness.

cs.LG↗

arXiv2020

Time for a Background Check! Uncovering the impact of Background Features on Deep Neural Networks

With increasing expressive power, deep neural networks have significantly improved the state-of-the-art on image classification datasets, such as ImageNet. In this paper, we investigate to what extent the increasing performance of deep neural networks is impacted by background features? In particular, we focus on background invariance, i.e., accuracy unaffected by switching background features and background influence, i.e., predictive power of background features itself when foreground is masked. We perform experiments with 32 different neural networks ranging from small-size networks to large-scale networks trained with up to one Billion images. Our investigations reveal that increasing expressive power of DNNs leads to higher influence of background features, while simultaneously, increases their ability to make the correct prediction when background features are removed or replaced with a randomly selected texture-based background.

cs.CV↗

arXiv2020

Programmable In-Network Obfuscation of Traffic

Recent advances in programmable switch hardware offer a fresh opportunity to protect user privacy. This paper presents PINOT, a lightweight in-network anonymity solution that runs at line rate within the memory and processing constraints of hardware switches. PINOT encrypts a client's IPv4 address with an efficient encryption scheme to hide the address from downstream ASes and the destination server. PINOT is readily deployable, requiring no end-user software or cooperation from networks other than the trusted network where it runs. We implement a PINOT prototype on the Barefoot Tofino switch, deploy PINOT in a campus network, and present results on protecting user identity against public DNS, NTP, and WireGuard VPN services.

cs.NI↗

arXiv2020

Grand Challenges in Resilience: Autonomous System Resilience through Design and Runtime Measures

A set of about 80 researchers, practitioners, and federal agency program managers participated in the NSF-sponsored Grand Challenges in Resilience Workshop held on Purdue campus on March 19-21, 2019. The workshop was divided into three themes: resilience in cyber, cyber-physical, and socio-technical systems. About 30 attendees in all participated in the discussions of cyber resilience. This article brings out the substantive parts of the challenges and solution approaches that were identified in the cyber resilience theme. In this article, we put forward the substantial challenges in cyber resilience in a few representative application domains and outline foundational solutions to address these challenges. These solutions fall into two broad themes: resilience-by-design and resilience-by-reaction. We use examples of autonomous systems as the application drivers motivating cyber resilience. We focus on some autonomous systems in the near horizon (autonomous ground and aerial vehicles) and also a little more distant (autonomous rescue and relief). For resilience-by-design, we focus on design methods in software that are needed for our cyber systems to be resilient. In contrast, for resilience-by-reaction, we discuss how to make systems resilient by responding, reconfiguring, or recovering at runtime when failures happen. We also discuss the notion of adaptive execution to improve resilience, execution transparently and adaptively among available execution platforms (mobile/embedded, edge, and cloud). For each of the two themes, we survey the current state, and the desired state and ways to get there. We conclude the paper by looking at the research challenges we will have to solve in the short and the mid-term to make the vision of resilient autonomous systems a reality.

cs.CR↗

arXiv2020

DP-Cryptography: Marrying Differential Privacy and Cryptography in Emerging Applications

Differential privacy (DP) has arisen as the state-of-the-art metric for quantifying individual privacy when sensitive data are analyzed, and it is starting to see practical deployment in organizations such as the US Census Bureau, Apple, Google, etc. There are two popular models for deploying differential privacy - standard differential privacy (SDP), where a trusted server aggregates all the data and runs the DP mechanisms, and local differential privacy (LDP), where each user perturbs their own data and perturbed data is analyzed. Due to security concerns arising from aggregating raw data at a single server, several real world deployments in industry have embraced the LDP model. However, systems based on the LDP model tend to have poor utility - "a gap" in the utility achieved as compared to systems based on the SDP model. In this work, we survey and synthesize emerging directions of research at the intersection of differential privacy and cryptography. First, we survey solutions that combine cryptographic primitives like secure computation and anonymous communication with differential privacy to give alternatives to the LDP model that avoid a trusted server as in SDP but close the gap in accuracy. These primitives introduce performance bottlenecks and necessitate efficient alternatives. Second, we synthesize work in an area we call "DP-Cryptography" - cryptographic primitives that are allowed to leak differentially private outputs. These primitives have orders of magnitude better performance than standard cryptographic primitives. DP-cryptographic primitives are perfectly suited for implementing alternatives to LDP, but are also applicable to scenarios where standard cryptographic primitives do not have practical implementations. Through this unique lens of research taxonomy, we survey ongoing research in these directions while also providing novel directions for future research.

cs.CR↗

arXiv2020

Querying Streaming System Monitoring Data for Enterprise System Anomaly Detection

The need for countering Advanced Persistent Threat (APT) attacks has led to the solutions that ubiquitously monitor system activities in each enterprise host, and perform timely abnormal system behavior detection over the stream of monitoring data. However, existing stream-based solutions lack explicit language constructs for expressing anomaly models that capture abnormal system behaviors, thus facing challenges in incorporating expert knowledge to perform timely anomaly detection over the large-scale monitoring data. To address these limitations, we build SAQL, a novel stream-based query system that takes as input, a real-time event feed aggregated from multiple hosts in an enterprise, and provides an anomaly query engine that queries the event feed to identify abnormal behaviors based on the specified anomaly models. SAQL provides a domain-specific query language, Stream-based Anomaly Query Language (SAQL), that uniquely integrates critical primitives for expressing major types of anomaly models. In the demo, we aim to show the complete usage scenario of SAQL by (1) performing an APT attack in a controlled environment, and (2) using SAQL to detect the abnormal behaviors in real time by querying the collected stream of system monitoring data that contains the attack traces. The audience will have the option to interact with the system and detect the attack footprints in real time via issuing queries and checking the query results through a command-line UI.

cs.CR↗