skip to main content

Title: CryptGPU: Fast Privacy-Preserving Machine Learning on the GPU
We introduce CryptGPU, a system for privacy-preserving machine learning that implements all operations on the GPU (graphics processing unit). Just as GPUs played a pivotal role in the success of modern deep learning, they are also essential for realizing scalable privacy-preserving deep learning. In this work, we start by introducing a new interface to losslessly embed cryptographic operations over secret-shared values (in a discrete domain) into floating-point operations that can be processed by highly-optimized CUDA kernels for linear algebra. We then identify a sequence of “GPU-friendly” cryptographic protocols to enable privacy-preserving evaluation of both linear and non-linear operations on the GPU. Our microbenchmarks indicate that our private GPU-based convolution protocol is over 150x faster than the analogous CPU-based protocol; for non-linear operations like the ReLU activation function, our GPU-based protocol is around 10x faster than its CPU analog. With CryptGPU, we support private inference and private training on convolutional neural networks with over 60 million parameters as well as handle large datasets like ImageNet. Compared to the previous state-of-the-art, when considering large models and datasets, our protocols achieve a 2x to 8x improvement in private inference and a 6x to 36x improvement for private training. Our work not only showcases the more » viability of performing secure multiparty computation (MPC) entirely on the GPU to enable fast privacy-preserving machine learning, but also highlights the importance of designing new MPC primitives that can take full advantage of the GPU's computing capabilities. « less
Authors:
; ; ;
Award ID(s):
2151131 1917414 2002985
Publication Date:
NSF-PAR ID:
10310226
Journal Name:
42nd IEEE Symposium on Security and Privacy
Sponsoring Org:
National Science Foundation
More Like this
  1. A protocol for two-party secure function evaluation (2P-SFE) aims to allow the parties to learn the output of function f of their private inputs, while leaking nothing more. In a sense, such a protocol realizes a trusted oracle that computes f and returns the result to both parties. There have been tremendous strides in efficiency over the past ten years, yet 2P-SFE protocols remain impractical for most real-time, online computations, particularly on modestly provisioned devices. Intel's Software Guard Extensions (SGX) provides hardware-protected execution environments, called enclaves, that may be viewed as trusted computation oracles. While SGX provides native CPU speed for secure computation, previous side-channel and micro-architecture attacks have demonstrated how security guarantees of enclaves can be compromised. In this paper, we explore a balanced approach to 2P-SFE on SGX-enabled processors by constructing a protocol for evaluating f relative to a partitioning of f. This approach alleviates the burden of trust on the enclave by allowing the protocol designer to choose which components should be evaluated within the enclave, and which via standard cryptographic techniques. We describe SGX-enabled SFE protocols (modeling the enclave as an oracle), and formalize the strongest-possible notion of 2P-SFE for our setting. We prove our protocolmore »meets this notion when properly realized. We implement the protocol and apply it to two practical problems: privacy-preserving queries to a database, and a version of Dijkstra's algorithm for privacy-preserving navigation. Our evaluation shows that our SGX-enabled SFE scheme enjoys a 38x increase in performance over garbled-circuit-based SFE. Finally, we justify modeling of the enclave as an oracle by implementing protections against known side-channels.« less
  2. Obeid, I. ; Selesnik, I. ; Picone, J. (Ed.)
    The Neuronix high-performance computing cluster allows us to conduct extensive machine learning experiments on big data [1]. This heterogeneous cluster uses innovative scheduling technology, Slurm [2], that manages a network of CPUs and graphics processing units (GPUs). The GPU farm consists of a variety of processors ranging from low-end consumer grade devices such as the Nvidia GTX 970 to higher-end devices such as the GeForce RTX 2080. These GPUs are essential to our research since they allow extremely compute-intensive deep learning tasks to be executed on massive data resources such as the TUH EEG Corpus [2]. We use TensorFlow [3] as the core machine learning library for our deep learning systems, and routinely employ multiple GPUs to accelerate the training process. Reproducible results are essential to machine learning research. Reproducibility in this context means the ability to replicate an existing experiment – performance metrics such as error rates should be identical and floating-point calculations should match closely. Three examples of ways we typically expect an experiment to be replicable are: (1) The same job run on the same processor should produce the same results each time it is run. (2) A job run on a CPU and GPU should producemore »identical results. (3) A job should produce comparable results if the data is presented in a different order. System optimization requires an ability to directly compare error rates for algorithms evaluated under comparable operating conditions. However, it is a difficult task to exactly reproduce the results for large, complex deep learning systems that often require more than a trillion calculations per experiment [5]. This is a fairly well-known issue and one we will explore in this poster. Researchers must be able to replicate results on a specific data set to establish the integrity of an implementation. They can then use that implementation as a baseline for comparison purposes. A lack of reproducibility makes it very difficult to debug algorithms and validate changes to the system. Equally important, since many results in deep learning research are dependent on the order in which the system is exposed to the data, the specific processors used, and even the order in which those processors are accessed, it becomes a challenging problem to compare two algorithms since each system must be individually optimized for a specific data set or processor. This is extremely time-consuming for algorithm research in which a single run often taxes a computing environment to its limits. Well-known techniques such as cross-validation [5,6] can be used to mitigate these effects, but this is also computationally expensive. These issues are further compounded by the fact that most deep learning algorithms are susceptible to the way computational noise propagates through the system. GPUs are particularly notorious for this because, in a clustered environment, it becomes more difficult to control which processors are used at various points in time. Another equally frustrating issue is that upgrades to the deep learning package, such as the transition from TensorFlow v1.9 to v1.13, can also result in large fluctuations in error rates when re-running the same experiment. Since TensorFlow is constantly updating functions to support GPU use, maintaining an historical archive of experimental results that can be used to calibrate algorithm research is quite a challenge. This makes it very difficult to optimize the system or select the best configurations. The overall impact of all of these issues described above is significant as error rates can fluctuate by as much as 25% due to these types of computational issues. Cross-validation is one technique used to mitigate this, but that is expensive since you need to do multiple runs over the data, which further taxes a computing infrastructure already running at max capacity. GPUs are preferred when training a large network since these systems train at least two orders of magnitude faster than CPUs [7]. Large-scale experiments are simply not feasible without using GPUs. However, there is a tradeoff to gain this performance. Since all our GPUs use the NVIDIA CUDA® Deep Neural Network library (cuDNN) [8], a GPU-accelerated library of primitives for deep neural networks, it adds an element of randomness into the experiment. When a GPU is used to train a network in TensorFlow, it automatically searches for a cuDNN implementation. NVIDIA’s cuDNN implementation provides algorithms that increase the performance and help the model train quicker, but they are non-deterministic algorithms [9,10]. Since our networks have many complex layers, there is no easy way to avoid this randomness. Instead of comparing each epoch, we compare the average performance of the experiment because it gives us a hint of how our model is performing per experiment, and if the changes we make are efficient. In this poster, we will discuss a variety of issues related to reproducibility and introduce ways we mitigate these effects. For example, TensorFlow uses a random number generator (RNG) which is not seeded by default. TensorFlow determines the initialization point and how certain functions execute using the RNG. The solution for this is seeding all the necessary components before training the model. This forces TensorFlow to use the same initialization point and sets how certain layers work (e.g., dropout layers). However, seeding all the RNGs will not guarantee a controlled experiment. Other variables can affect the outcome of the experiment such as training using GPUs, allowing multi-threading on CPUs, using certain layers, etc. To mitigate our problems with reproducibility, we first make sure that the data is processed in the same order during training. Therefore, we save the data from the last experiment and to make sure the newer experiment follows the same order. If we allow the data to be shuffled, it can affect the performance due to how the model was exposed to the data. We also specify the float data type to be 32-bit since Python defaults to 64-bit. We try to avoid using 64-bit precision because the numbers produced by a GPU can vary significantly depending on the GPU architecture [11-13]. Controlling precision somewhat reduces differences due to computational noise even though technically it increases the amount of computational noise. We are currently developing more advanced techniques for preserving the efficiency of our training process while also maintaining the ability to reproduce models. In our poster presentation we will demonstrate these issues using some novel visualization tools, present several examples of the extent to which these issues influence research results on electroencephalography (EEG) and digital pathology experiments and introduce new ways to manage such computational issues.« less
  3. Canteaut, Anne ; Standaert, Francois-Xavier (Ed.)
    Secure multi-party computation (MPC) allows multiple par-ties to perform secure joint computations on their private inputs. To-day, applications for MPC are growing with thousands of parties wish-ing to build federated machine learning models or trusted setups for blockchains. To address such scenarios we propose a suite of novel MPC protocols that maximize throughput when run with large numbers of parties. In particular, our protocols have both communication and computation complexity that decrease with the number of parties. Our protocols build on prior protocols based on packed secret-sharing, introducing new techniques to build more efficient computation for general circuits. Specifically, we introduce a new approach for handling linear attacks that arise in protocols using packed secret-sharing and we propose a method for unpacking shared multiplication triples without increasingthe asymptotic costs. Compared with prior work, we avoid the log|C|overhead required when generically compiling circuits of size |C| for use in a SIMD computation, and we improve over folklore “committee-based” solutions by a factor of O(s), the statistical security parameter. In practice, our protocol is up to 10X faster than any known construction, under a reasonable set of parameters.
  4. Canteaut, Anne ; Standaert, Francois-Xavier (Ed.)
    Secure multi-party computation (MPC) allows multiple par-ties to perform secure joint computations on their private inputs. To-day, applications for MPC are growing with thousands of parties wish-ing to build federated machine learning models or trusted setups for blockchains. To address such scenarios we propose a suite of novel MPC protocols that maximize throughput when run with large numbers of parties. In particular, our protocols have both communication and computation complexity that decrease with the number of parties. Our protocols build on prior protocols based on packed secret-sharing, introducing new techniques to build more efficient computation for general circuits. Specifically, we introduce a new approach for handling linear attacks that arise in protocols using packed secret-sharing and we propose a method for unpacking shared multiplication triples without increasingthe asymptotic costs. Compared with prior work, we avoid the log|C|overhead required when generically compiling circuits of size |C| for use in a SIMD computation, and we improve over folklore “committee-based” solutions by a factor of O(s), the statistical security parameter. In practice, our protocol is up to 10X faster than any known construction, under a reasonable set of parameters.
  5. New advances in machine learning have made Automated Speech Recognition (ASR) systems practical and more scalable. These systems, however, pose serious privacy threats as speech is a rich source of sensitive acoustic and textual information. Although offline and open-source ASR eliminates the privacy risks, its transcription performance is inferior to that of cloud-based ASR systems, especially for real-world use cases. In this paper, we propose Prεεch, an end-to-end speech transcription system which lies at an intermediate point in the privacy-utility spectrum. It protects the acoustic features of the speakers’ voices and protects the privacy of the textual content at an improved performance relative to offline ASR. Additionally, Prεεch provides several control knobs to allow customizable utility-usability-privacy trade-off. It relies on cloud-based services to transcribe a speech file after applying a series of privacy-preserving operations on the user’s side. We perform a comprehensive evaluation of Prεεch, using diverse real-world datasets, that demonstrates its effectiveness. Prεεch provides transcription at a 2% to 32.25% (mean 17.34%) relative improvement in word error rate over Deep Speech, while fully obfuscating the speakers' voice biometrics and allowing only a differentially private view of the textual content.