Sharing real-world speech utterances is key to the training and deployment of voice-based services. However, it also raises privacy risks as speech contains a wealth of personal data. Speaker anonymization aims to remove speaker information from a speech utterance while leaving its linguistic and prosodic attributes intact. State-of-the-art techniques operate by disentangling the speaker information (represented via a speaker embedding) from these attributes and re-synthesizing speech based on the speaker embedding of another speaker. Prior research in the privacy community has shown that anonymization often provides brittle privacy protection, even less so any provable guarantee. In this work, we show that disentanglement is indeed not perfect: linguistic and prosodic attributes still contain speaker information. We remove speaker information from these attributes by introducing differentially private feature extractors based on an autoencoder and an automatic speech recognizer, respectively, trained using noise layers. We plug these extractors in the state-of-the-art anonymization pipeline and generate, for the first time, differentially private utterances with a provable upper bound on the speaker information they contain. We evaluate empirically the privacy and utility resulting from our differentially private speaker anonymization approach on the LibriSpeech data set. Experimental results show that the generated utterances retain very high utility for automatic speech recognition training and inference, while being much better protected against strong adversaries who leverage the full knowledge of the anonymization process to try to infer the speaker identity.
Recent years have seen a surge of popularity of acoustics-enabled personal devices powered by machine learning. Yet, machine learning has proven to be vulnerable to adversarial examples. Large number of modern systems protect themselves against such attacks by targeting the artificiality, i.e., they deploy mechanisms to detect the lack of human involvement in generating the adversarial examples. However, these defenses implicitly assume that humans are incapable of producing meaningful and targeted adversarial examples. In this paper, we show that this base assumption is wrong. In particular, we demonstrate that for tasks like speaker identification, a human is capable of producing analog adversarial examples directly with little cost and supervision: by simply speaking through a tube, an adversary reliably impersonates other speakers in eyes of ML models for speaker identification. Our findings extend to a range of other acoustic-biometric tasks such as liveness, bringing into question their use in security-critical settings in real life, such as phone banking.
In model extraction attacks, adversaries can steal a machine learning model exposed via a public API by repeatedly querying it and adjusting their own model based on obtained predictions. To prevent model stealing, existing defenses focus on detecting malicious queries, truncating, or distorting outputs, thus necessarily introducing a tradeoff between robustness and model utility for legitimate users. Instead, we propose to impede model extraction by requiring users to complete a proof-of-work before they can read the model's predictions. This deters attackers by greatly increasing (even up to 100x) the computational effort needed to leverage query access for model extraction. Since we calibrate the effort required to complete the proof-of-work to each query, this only introduces a slight overhead for regular users (up to 2x). To achieve this, our calibration applies tools from differential privacy to measure the information revealed by a query. Our method requires no modification of the victim model and can be applied by machine learning practitioners to guard their publicly exposed models against being easily stolen.
In federated learning (FL), data does not leave personal devices when they are jointly training a machine learning model. Instead, these devices share gradients with a central party (e.g., a company). Because data never "leaves" personal devices, FL is presented as privacy-preserving. Yet, recently it was shown that this protection is but a thin facade, as even a passive attacker observing gradients can reconstruct data of individual users. In this paper, we argue that prior work still largely underestimates the vulnerability of FL. This is because prior efforts exclusively consider passive attackers that are honest-but-curious. Instead, we introduce an active and dishonest attacker acting as the central party, who is able to modify the shared model's weights before users compute model gradients. We call the modified weights "trap weights". Our active attacker is able to recover user data perfectly and at near zero costs: the attack requires no complex optimization objectives. Instead, it exploits inherent data leakage from model gradients and amplifies this effect by maliciously altering the weights of the shared model. These specificities enable our attack to scale to models trained with large mini-batches of data. Where attackers from prior work require hours to recover a single data point, our method needs milliseconds to capture the full mini-batch of data from both fully-connected and convolutional deep neural networks. Finally, we consider mitigations. We observe that current implementations of differential privacy (DP) in FL are flawed, as they explicitly trust the central party with the crucial task of adding DP noise, and thus provide no protection against a malicious central party. We also consider other defenses and explain why they are similarly inadequate. A significant redesign of FL is required for it to provide any meaningful form of data privacy to users.
Machine unlearning, i.e. having a model forget about some of its training data, has become increasingly more important as privacy legislation promotes variants of the right-to-be-forgotten. In the context of deep learning, approaches for machine unlearning are broadly categorized into two classes: exact unlearning methods, where an entity has formally removed the data point's impact on the model by retraining the model from scratch, and approximate unlearning, where an entity approximates the model parameters one would obtain by exact unlearning to save on compute costs. In this paper we first show that the definition that underlies approximate unlearning, which seeks to prove the approximately unlearned model is close to an exactly retrained model, is incorrect because one can obtain the same model using different datasets. Thus one could unlearn without modifying the model at all. We then turn to exact unlearning approaches and ask how to verify their claims of unlearning. Our results show that even for a given training trajectory one cannot formally prove the absence of certain data points used during training. We thus conclude that unlearning is only well-defined at the algorithmic level, where an entity's only possible auditable claim to unlearning is that they used a particular algorithm designed to allow for external scrutiny during an audit.
For many differentially private algorithms, such as the prominent noisy stochastic gradient descent (DP-SGD), the analysis needed to bound the privacy leakage of a single training run is well understood. However, few studies have reasoned about the privacy leakage resulting from the multiple training runs needed to fine tune the value of the training algorithm's hyperparameters. In this work, we first illustrate how simply setting hyperparameters based on non-private training runs can leak private information. Motivated by this observation, we then provide privacy guarantees for hyperparameter search procedures within the framework of Renyi Differential Privacy. Our results improve and extend the work of Liu and Talwar (STOC 2019). Our analysis supports our previous observation that tuning hyperparameters does indeed leak private information, but we prove that, under certain assumptions, this leakage is modest, as long as each candidate training run needed to select hyperparameters is itself differentially private.
Machine unlearning is the process through which a deployed machine learning model forgets about one of its training data points. While naively retraining the model from scratch is an option, it is almost always associated with a large computational effort for deep learning models. Thus, several approaches to approximately unlearn have been proposed along with corresponding metrics that formalize what it means for a model to forget about a data point. In this work, we first taxonomize approaches and metrics of approximate unlearning. As a result, we identify verification error, i.e., the L2 difference between the weights of an approximately unlearned and a naively retrained model, as a metric approximate unlearning should optimize for as it implies a large class of other metrics. We theoretically analyze the canonical stochastic gradient descent (SGD) training algorithm to surface the variables which are relevant to reducing the verification error of approximate unlearning for SGD. From this analysis, we first derive an easy-to-compute proxy for verification error (termed unlearning error). The analysis also informs the design of a new training objective penalty that limits the overall change in weights during SGD and as a result facilitates approximate unlearning with lower verification error. We validate our theoretical work through an empirical evaluation on CIFAR-10, CIFAR-100, and IMDB sentiment analysis.
Sophisticated machine learning (ML) models to inform trading in the financial sector create problems of interpretability and risk management. Seemingly robust forecasting models may behave erroneously in out of distribution settings. In 2020, some of the world's most sophisticated quant hedge funds suffered losses as their ML models were first underhedged, and then overcompensated. We implement a gradient-based approach for precisely stress-testing how a trading model's forecasts can be manipulated, and their effects on downstream tasks at the trading execution level. We construct inputs -- whether in changes to sentiment or market variables -- that efficiently affect changes in the return distribution. In an industry-standard trading pipeline, we perturb model inputs for eight S&P 500 stocks. We find our approach discovers seemingly in-sample input settings that result in large negative shifts in return distributions. We provide the financial community with mechanisms to interpret ML forecasts in trading systems. For the security community, we provide a compelling application where studying ML robustness necessitates that one capture an end-to-end system's performance rather than study a ML model in isolation. Indeed, we show in our evaluation that errors in the forecasting model's predictions alone are not sufficient for trading decisions made based on these forecasts to yield a negative return.
The application of machine learning (ML) in computer systems introduces not only many benefits but also risks to society. In this paper, we develop the concept of ML governance to balance such benefits and risks, with the aim of achieving responsible applications of ML. Our approach first systematizes research towards ascertaining ownership of data and models, thus fostering a notion of identity specific to ML systems. Building on this foundation, we use identities to hold principals accountable for failures of ML systems through both attribution and auditing. To increase trust in ML systems, we then survey techniques for developing assurance, i.e., confidence that the system meets its security requirements and does not exhibit certain known failures. This leads us to highlight the need for techniques that allow a model owner to manage the life cycle of their system, e.g., to patch or retire their ML system. Put altogether, our systematization of knowledge standardizes the interactions between principals involved in the deployment of ML throughout its life cycle. We highlight opportunities for future work, e.g., to formalize the resulting game between ML principals.
Machine learning (ML) models are known to be vulnerable to adversarial examples. Applications of ML to voice biometrics authentication are no exception. Yet, the implications of audio adversarial examples on these real-world systems remain poorly understood given that most research targets limited defenders who can only listen to the audio samples. Conflating detectability of an attack with human perceptibility, research has focused on methods that aim to produce imperceptible adversarial examples which humans cannot distinguish from the corresponding benign samples. We argue that this perspective is coarse for two reasons: 1. Imperceptibility is impossible to verify; it would require an experimental process that encompasses variations in listener training, equipment, volume, ear sensitivity, types of background noise etc, and 2. It disregards pipeline-based detection clues that realistic defenders leverage. This results in adversarial examples that are ineffective in the presence of knowledgeable defenders. Thus, an adversary only needs an audio sample to be plausible to a human. We thus introduce surreptitious adversarial examples, a new class of attacks that evades both human and pipeline controls. In the white-box setting, we instantiate this class with a joint, multi-stage optimization attack. Using an Amazon Mechanical Turk user study, we show that this attack produces audio samples that are more surreptitious than previous attacks that aim solely for imperceptibility. Lastly we show that surreptitious adversarial examples are challenging to develop in the black-box setting.