Indranil Gupta

Beomyeol Jeon, Linda Cai, Chirag Shetty et al.

Machine Learning graphs (or models) can be challenging or impossible to train when either devices have limited memory, or models are large. To split the model across devices, learning-based approaches are still popular. While these result in model placements that train fast on data (i.e., low step times), learning-based model-parallelism is time-consuming, taking many hours or days to create a placement plan of operators on devices. We present the Baechi system, the first to adopt an algorithmic approach to the placement problem for running machine learning training graphs on small clusters of memory-constrained devices. We integrate our implementation of Baechi into two popular open-source learning frameworks: TensorFlow and PyTorch. Our experimental results using GPUs show that: (i) Baechi generates placement plans 654 X - 206K X faster than state-of-the-art learning-based approaches, and (ii) Baechi-placed model's step (training) time is comparable to expert placements in PyTorch, and only up to 6.2% worse than expert placements in TensorFlow. We prove mathematically that our two algorithms are within a constant factor of the optimal. Our work shows that compared to learning-based approaches, algorithmic approaches can face different challenges for adaptation to Machine learning systems, but also they offer proven bounds, and significant performance benefits.

Maleeha Masood, Shreya Kannan, Om Chabra et al.

Short video platforms like TikTok, Instagram Reels, and YouTube Shorts have gained immense popularity in the last few years and are responsible for a large and growing fraction of Internet traffic. We identify two unique opportunities for improving short video delivery using their existing interactions with content delivery networks (CDNs). First, short videos use a push-based recommendation system, where the user is presented a sequence of videos recommended by the algorithm rather than user explicitly picking content to watch (e.g., in YouTube). Such push-based short video systems offer a unique opportunity for system design by providing visibility into upcoming requests. Second, the popularity of these videos follows a highly skewed Pareto distribution, leading to geographical and temporal overlap amongst videos being served. We leverage these opportunities to build SILC - a lookahead-aware caching system, aimed at (i) reducing CDN cache miss rates, as well as (ii) reducing midgress bandwidth between the CDN and the origin server. Our evaluation of SILC uses traces that we collect from real users, through (i) an in-person user study, and (ii) a data donation program involving 100 TikTok users across the world. Using a combination of these traces, we simulate traffic from 10,000 simultaneous users. Our evaluation shows that, compared to 10 state-of-the-art heuristic and learning-based cache eviction policies, SILC reduces a CDN's midgress costs by 11.1% to 111%.

Cong Xie, Shuai Zheng, Oluwasanmi Koyejo et al.

The scalability of Distributed Stochastic Gradient Descent (SGD) is today limited by communication bottlenecks. We propose a novel SGD variant: Communication-efficient SGD with Error Reset, or CSER. The key idea in CSER is first a new technique called "error reset" that adapts arbitrary compressors for SGD, producing bifurcated local models with periodic reset of resulting local residual errors. Second we introduce partial synchronization for both the gradients and the models, leveraging advantages from them. We prove the convergence of CSER for smooth non-convex problems. Empirical results show that when combined with highly aggressive compressors, the CSER algorithms accelerate the distributed training by nearly 10x for CIFAR-100, and by 4.5x for ImageNet.

Cong Xie, Oluwasanmi Koyejo, Indranil Gupta et al.

When scaling distributed training, the communication overhead is often the bottleneck. In this paper, we propose a novel SGD variant with reduced communication and adaptive learning rates. We prove the convergence of the proposed algorithm for smooth but non-convex problems. Empirical results show that the proposed algorithm significantly reduces the communication overhead, which, in turn, reduces the training time by up to 30% for the 1B word dataset.

Cong Xie, Sanmi Koyejo, Indranil Gupta

We propose Zeno++, a new robust asynchronous Stochastic Gradient Descent~(SGD) procedure which tolerates Byzantine failures of the workers. In contrast to previous work, Zeno++ removes some unrealistic restrictions on worker-server communications, allowing for fully asynchronous updates from anonymous workers, arbitrarily stale worker updates, and the possibility of an unbounded number of Byzantine workers. The key idea is to estimate the descent of the loss value after the candidate gradient is applied, where large descent values indicate that the update results in optimization progress. We prove the convergence of Zeno++ for non-convex problems under Byzantine failures. Experimental results show that Zeno++ outperforms existing approaches.

Cong Xie, Sanmi Koyejo, Indranil Gupta

We consider distributed on-device learning with limited communication and security requirements. We propose a new robust distributed optimization algorithm with efficient communication and attack tolerance. The proposed algorithm has provable convergence and robustness under non-IID settings. Empirical results show that the proposed algorithm stabilizes the convergence and tolerates data poisoning on a small number of workers.

Cong Xie, Sanmi Koyejo, Indranil Gupta

Recently, new defense techniques have been developed to tolerate Byzantine failures for distributed machine learning. The Byzantine model captures workers that behave arbitrarily, including malicious and compromised workers. In this paper, we break two prevailing Byzantine-tolerant techniques. Specifically we show robust aggregation methods for synchronous SGD -- coordinate-wise median and Krum -- can be broken using new attack strategies based on inner product manipulation. We prove our results theoretically, as well as show empirical validation.

Cong Xie, Sanmi Koyejo, Indranil Gupta

Federated learning enables training on a massive number of edge devices. To improve flexibility and scalability, we propose a new asynchronous federated optimization algorithm. We prove that the proposed approach has near-linear convergence to a global optimum, for both strongly convex and a restricted family of non-convex problems. Empirical results show that the proposed algorithm converges quickly and tolerates staleness in various applications.

Cong Xie, Oluwasanmi Koyejo, Indranil Gupta

We present Zeno, a technique to make distributed machine learning, particularly Stochastic Gradient Descent (SGD), tolerant to an arbitrary number of faulty workers. Zeno generalizes previous results that assumed a majority of non-faulty nodes; we need assume only one non-faulty worker. Our key idea is to suspect workers that are potentially defective. Since this is likely to lead to false positives, we use a ranking-based preference mechanism. We prove the convergence of SGD for non-convex problems under these scenarios. Experimental results show that Zeno outperforms existing approaches.

Cong Xie, Oluwasanmi Koyejo, Indranil Gupta

We propose a novel robust aggregation rule for distributed synchronous Stochastic Gradient Descent~(SGD) under a general Byzantine failure model. The attackers can arbitrarily manipulate the data transferred between the servers and the workers in the parameter server~(PS) architecture. We prove the Byzantine resilience of the proposed aggregation rules. Empirical analysis shows that the proposed techniques outperform current approaches for realistic use cases and Byzantine attack scenarios.

Cong Xie, Oluwasanmi Koyejo, Indranil Gupta

We propose three new robust aggregation rules for distributed synchronous Stochastic Gradient Descent~(SGD) under a general Byzantine failure model. The attackers can arbitrarily manipulate the data transferred between the servers and the workers in the parameter server~(PS) architecture. We prove the Byzantine resilience properties of these aggregation rules. Empirical analysis shows that the proposed techniques outperform current approaches for realistic use cases and Byzantine attack scenarios.