Neural Information Processing Systems http://nips.cc/

When workers cannot form a fully connected communication topology or the communication latency is high (e.g., in

Optimizing distributed learning systems is an art of balancing between computation and communication. There have been two lines of research that try to deal with slower networks: {\em communication compression} for low bandwidth networks, and {\em decentralization} for high latency networks. In this paper, We explore a natural question: {\em can the combination of both techniques lead to a system that is robust to both bandwidth and latency?} Although the system implication of such combination is trivial, the underlying theoretical principle and algorithm design is challenging: unlike centralized algorithms, simply compressing {\rc exchanged information, even in an unbiased stochastic way, within the decentralized network would accumulate the error and cause divergence.} In this paper, we develop a framework of quantized, decentralized training and propose two different strategies, which we call {\em extrapolation compression} and {\em difference compression}. We analyze both algorithms and prove both converge at the rate of $O(1/\sqrt{nT})$ where $n$ is the number of workers and $T$ is the number of iterations, matching the convergence rate for full precision, centralized training. We validate our algorithms and find that our proposed algorithm outperforms the best of merely decentralized and merely quantized algorithm significantly for networks with {\em both} high latency and low bandwidth.

Optimizing distributed learning systems is an art of balancing between computation and communication. There have been two lines of research that try to deal with slower networks: {\em communication compression} for low bandwidth networks, and {\em decentralization} for high latency networks. In this paper, We explore a natural question: {\em can the combination of both techniques lead to a system that is robust to both bandwidth and latency?} Although the system implication of such combination is trivial, the underlying theoretical principle and algorithm design is challenging: unlike centralized algorithms, simply compressing {\rc exchanged information, even in an unbiased stochastic way, within the decentralized network would accumulate the error and cause divergence.} In this paper, we develop a framework of quantized, decentralized training and propose two different strategies, which we call {\em extrapolation compression} and {\em difference compression}. We analyze both algorithms and prove both converge at the rate of $O(1/\sqrt{nT})$ where $n$ is the number of workers and $T$ is the number of iterations, matching the convergence rate for full precision, centralized training. We validate our algorithms and find that our proposed algorithm outperforms the best of merely decentralized and merely quantized algorithm significantly for networks with {\em both} high latency and low bandwidth.

While training a machine learning model using multiple workers, each of which collects data from their own data sources, it would be most useful when the data collected from different workers can be {\em unique} and {\em different}. Ironically, recent analysis of decentralized parallel stochastic gradient descent (D-PSGD) relies on the assumption that the data hosted on different workers are {\em not too different}. In this paper, we ask the question: {\em Can we design a decentralized parallel stochastic gradient descent algorithm that is less sensitive to the data variance across workers?} In this paper, we present D$^2$, a novel decentralized parallel stochastic gradient descent algorithm designed for large data variance \xr{among workers} (imprecisely, "decentralized" data). The core of D$^2$ is a variance blackuction extension of the standard D-PSGD algorithm, which improves the convergence rate from $O\left({\sigma \over \sqrt{nT}} + {(n\zeta^2)^{\frac{1}{3}} \over T^{2/3}}\right)$ to $O\left({\sigma \over \sqrt{nT}}\right)$ where $\zeta^{2}$ denotes the variance among data on different workers. As a result, D$^2$ is robust to data variance among workers. We empirically evaluated D$^2$ on image classification tasks where each worker has access to only the data of a limited set of labels, and find that D$^2$ significantly outperforms D-PSGD.

Optimizing distributed learning systems is an art of balancing between computation and communication. There have been two lines of research that try to deal with slower networks: {\em quantization} for low bandwidth networks, and {\em decentralization} for high latency networks. In this paper, we explore a natural question: {\em can the combination of both decentralization and quantization lead to a system that is robust to both bandwidth and latency?} Although the system implication of such combination is trivial, the underlying theoretical principle and algorithm design is challenging: simply quantizing data sent in a decentralized training algorithm would accumulate the error. In this paper, we develop a framework of quantized, decentralized training and propose two different strategies, which we call {\em extrapolation compression} and {\em difference compression}. We analyze both algorithms and prove both converge at the rate of $O(1/\sqrt{nT})$ where $n$ is the number of workers and $T$ is the number of iterations, matching the {\rc convergence} rate for full precision, centralized training. We evaluate our algorithms on training deep learning models, and find that our proposed algorithm outperforms the best of merely decentralized and merely quantized algorithm significantly for networks with {\em both} high latency and low bandwidth.

Stochastic gradient descent (SGD) is a well known method for regression and classification tasks. However, it is an inherently sequential algorithm at each step, the processing of the current example depends on the parameters learned from the previous examples. Prior approaches to parallelizing linear learners using SGD, such as HOGWILD! and ALLREDUCE, do not honor these dependencies across threads and thus can potentially suffer poor convergence rates and/or poor scalability. This paper proposes SYMSGD, a parallel SGD algorithm that, to a first-order approximation, retains the sequential semantics of SGD. Each thread learns a local model in addition to a model combiner, which allows local models to be combined to produce the same result as what a sequential SGD would have produced. This paper evaluates SYMSGD's accuracy and performance on 6 datasets on a shared-memory machine shows upto 11x speedup over our heavily optimized sequential baseline on 16 cores and 2.2x, on average, faster than HOGWILD!.