arXiv subjects

Xiaoyun Li

Publications and source records attributed to Xiaoyun Li.

At least 19 recordsLinked to original sources

arXiv2025

DaiFu: In-Situ Crash Recovery for Deep Learning Systems

Deep learning (DL) systems have been widely adopted in many areas, and are becoming even more popular with the emergence of large language models. However, due to the complex software stacks involved in their development and execution, crashes are unavoidable and common. Crashes severely waste computing resources and hinder development productivity, so efficient crash recovery is crucial. Existing solutions, such as checkpoint-retry, are too heavyweight for fast recovery from crashes caused by minor programming errors or transient runtime errors. Therefore, we present DaiFu, an in-situ recovery framework for DL systems. Through a lightweight code transformation to a given DL system, DaiFu augments it to intercept crashes in situ and enables dynamic and instant updates to its program running context (e.g., code, configurations, and other data) for agile crash recovery. Our evaluation shows that DaiFu helps reduce the restore time for crash recovery, achieving a 1372x speedup compared with state-of-the-art solutions. Meanwhile, the overhead of DaiFu is negligible (under 0.40%). We also construct a benchmark spanning 7 distinct crash scenarios in DL systems, and show the effectiveness of DaiFu in diverse situations.

cs.SE↗

arXiv2024

Stochastic Controlled Averaging for Federated Learning with Communication Compression

Communication compression, a technique aiming to reduce the information volume to be transmitted over the air, has gained great interests in Federated Learning (FL) for the potential of alleviating its communication overhead. However, communication compression brings forth new challenges in FL due to the interplay of compression-incurred information distortion and inherent characteristics of FL such as partial participation and data heterogeneity. Despite the recent development, the performance of compressed FL approaches has not been fully exploited. The existing approaches either cannot accommodate arbitrary data heterogeneity or partial participation, or require stringent conditions on compression. In this paper, we revisit the seminal stochastic controlled averaging method by proposing an equivalent but more efficient/simplified formulation with halved uplink communication costs. Building upon this implementation, we propose two compressed FL algorithms, SCALLION and SCAFCOM, to support unbiased and biased compression, respectively. Both the proposed methods outperform the existing compressed FL methods in terms of communication and computation complexities. Moreover, SCALLION and SCAFCOM accommodates arbitrary data heterogeneity and do not make any additional assumptions on compression errors. Experiments show that SCALLION and SCAFCOM can match the performance of corresponding full-precision FL approaches with substantially reduced uplink communication, and outperform recent compressed FL methods under the same communication budget.

math.OC↗

arXiv2023

LogShrink: Effective Log Compression by Leveraging Commonality and Variability of Log Data

Log data is a crucial resource for recording system events and states during system execution. However, as systems grow in scale, log data generation has become increasingly explosive, leading to an expensive overhead on log storage, such as several petabytes per day in production. To address this issue, log compression has become a crucial task in reducing disk storage while allowing for further log analysis. Unfortunately, existing general-purpose and log-specific compression methods have been limited in their ability to utilize log data characteristics. To overcome these limitations, we conduct an empirical study and obtain three major observations on the characteristics of log data that can facilitate the log compression task. Based on these observations, we propose LogShrink, a novel and effective log compression method by leveraging commonality and variability of log data. An analyzer based on longest common subsequence and entropy techniques is proposed to identify the latent commonality and variability in log messages. The key idea behind this is that the commonality and variability can be exploited to shrink log data with a shorter representation. Besides, a clustering-based sequence sampler is introduced to accelerate the commonality and variability analyzer. The extensive experimental results demonstrate that LogShrink can exceed baselines in compression ratio by 16% to 356% on average while preserving a reasonable compression speed.

cs.SE↗

arXiv2023

CSPM: A Contrastive Spatiotemporal Preference Model for CTR Prediction in On-Demand Food Delivery Services

Click-through rate (CTR) prediction is a crucial task in the context of an online on-demand food delivery (OFD) platform for precisely estimating the probability of a user clicking on food items. Unlike universal e-commerce platforms such as Taobao and Amazon, user behaviors and interests on the OFD platform are more location and time-sensitive due to limited delivery ranges and regional commodity supplies. However, existing CTR prediction algorithms in OFD scenarios concentrate on capturing interest from historical behavior sequences, which fails to effectively model the complex spatiotemporal information within features, leading to poor performance. To address this challenge, this paper introduces the Contrastive Sres under different search states using three modules: contrastive spatiotemporal representation learning (CSRL), spatiotemporal preference extractor (StPE), and spatiotemporal information filter (StIF). CSRL utilizes a contrastive learning framework to generate a spatiotemporal activation representation (SAR) for the search action. StPE employs SAR to activate users' diverse preferences related to location and time from the historical behavior sequence field, using a multi-head attention mechanism. StIF incorporates SAR into a gating network to automatically capture important features with latent spatiotemporal effects. Extensive experiments conducted on two large-scale industrial datasets demonstrate the state-of-the-art performance of CSPM. Notably, CSPM has been successfully deployed in Alibaba's online OFD platform Ele.me, resulting in a significant 0.88% lift in CTR, which has substantial business implications.

cs.IR↗

arXiv2023

Differential Privacy with Random Projections and Sign Random Projections

In this paper, we develop a series of differential privacy (DP) algorithms from a family of random projections (RP) for general applications in machine learning, data mining, and information retrieval. Among the presented algorithms, iDP-SignRP is remarkably effective under the setting of ``individual differential privacy'' (iDP), based on sign random projections (SignRP). Also, DP-SignOPORP considerably improves existing algorithms in the literature under the standard DP setting, using ``one permutation + one random projection'' (OPORP), where OPORP is a variant of the celebrated count-sketch method with fixed-length binning and normalization. Without taking signs, among the DP-RP family, DP-OPORP achieves the best performance. Our key idea for improving DP-RP is to take only the signs, i.e., $sign(x_j) = sign\left(\sum_{i=1}^p u_i w_{ij}\right)$, of the projected data. The intuition is that the signs often remain unchanged when the original data ($u$) exhibit small changes (according to the ``neighbor'' definition in DP). In other words, the aggregation and quantization operations themselves provide good privacy protections. We develop a technique called ``smooth flipping probability'' that incorporates this intuitive privacy benefit of SignRPs and improves the standard DP bit flipping strategy. Based on this technique, we propose DP-SignOPORP which satisfies strict DP and outperforms other DP variants based on SignRP (and RP), especially when $ε$ is not very large (e.g., $ε= 5\sim10$). Moreover, if an application scenario accepts individual DP, then we immediately obtain an algorithm named iDP-SignRP which achieves excellent utilities even at small~$ε$ (e.g., $ε<0.5$).

cs.CR↗

arXiv2023

Differentially Private One Permutation Hashing and Bin-wise Consistent Weighted Sampling

Minwise hashing (MinHash) is a standard algorithm widely used in the industry, for large-scale search and learning applications with the binary (0/1) Jaccard similarity. One common use of MinHash is for processing massive n-gram text representations so that practitioners do not have to materialize the original data (which would be prohibitive). Another popular use of MinHash is for building hash tables to enable sub-linear time approximate near neighbor (ANN) search. MinHash has also been used as a tool for building large-scale machine learning systems. The standard implementation of MinHash requires applying $K$ random permutations. In comparison, the method of one permutation hashing (OPH), is an efficient alternative of MinHash which splits the data vectors into $K$ bins and generates hash values within each bin. OPH is substantially more efficient and also more convenient to use. In this paper, we combine the differential privacy (DP) with OPH (as well as MinHash), to propose the DP-OPH framework with three variants: DP-OPH-fix, DP-OPH-re and DP-OPH-rand, depending on which densification strategy is adopted to deal with empty bins in OPH. A detailed roadmap to the algorithm design is presented along with the privacy analysis. An analytical comparison of our proposed DP-OPH methods with the DP minwise hashing (DP-MH) is provided to justify the advantage of DP-OPH. Experiments on similarity search confirm the merits of DP-OPH, and guide the choice of the proper variant in different practical scenarios. Our technique is also extended to bin-wise consistent weighted sampling (BCWS) to develop a new DP algorithm called DP-BCWS for non-binary data. Experiments on classification tasks demonstrate that DP-BCWS is able to achieve excellent utility at around $ε= 5\sim 10$, where $ε$ is the standard parameter in the language of $(ε, δ)$-DP.

stat.ML↗

arXiv2023

OPORP: One Permutation + One Random Projection

Consider two $D$-dimensional data vectors (e.g., embeddings): $u, v$. In many embedding-based retrieval (EBR) applications where the vectors are generated from trained models, $D=256\sim 1024$ are common. In this paper, OPORP (one permutation + one random projection) uses a variant of the ``count-sketch'' type of data structures for achieving data reduction/compression. With OPORP, we first apply a permutation on the data vectors. A random vector $r$ is generated i.i.d. with moments: $E(r_i) = 0, E(r_i^2)=1, E(r_i^3) =0, E(r_i^4)=s$. We multiply (as dot product) $r$ with all permuted data vectors. Then we break the $D$ columns into $k$ equal-length bins and aggregate (i.e., sum) the values in each bin to obtain $k$ samples from each data vector. One crucial step is to normalize the $k$ samples to the unit $l_2$ norm. We show that the estimation variance is essentially: $(s-1)A + \frac{D-k}{D-1}\frac{1}{k}\left[ (1-ρ^2)^2 -2A\right]$, where $A\geq 0$ is a function of the data ($u,v$). This formula reveals several key properties: (1) We need $s=1$. (2) The factor $\frac{D-k}{D-1}$ can be highly beneficial in reducing variances. (3) The term $\frac{1}{k}(1-ρ^2)^2$ is a substantial improvement compared with $\frac{1}{k}(1+ρ^2)$, which corresponds to the un-normalized estimator. We illustrate that by letting the $k$ in OPORP to be $k=1$ and repeat the procedure $m$ times, we exactly recover the work of ``very spars random projections'' (VSRP). This immediately leads to a normalized estimator for VSRP which substantially improves the original estimator of VSRP. In summary, with OPORP, the two key steps: (i) the normalization and (ii) the fixed-length binning scheme, have considerably improved the accuracy in estimating the cosine similarity, which is a routine (and crucial) task in modern embedding-based retrieval (EBR) applications.

stat.ML↗

arXiv2023

Building K-Anonymous User Cohorts with\\ Consecutive Consistent Weighted Sampling (CCWS)

To retrieve personalized campaigns and creatives while protecting user privacy, digital advertising is shifting from member-based identity to cohort-based identity. Under such identity regime, an accurate and efficient cohort building algorithm is desired to group users with similar characteristics. In this paper, we propose a scalable $K$-anonymous cohort building algorithm called {\em consecutive consistent weighted sampling} (CCWS). The proposed method combines the spirit of the ($p$-powered) consistent weighted sampling and hierarchical clustering, so that the $K$-anonymity is ensured by enforcing a lower bound on the size of cohorts. Evaluations on a LinkedIn dataset consisting of $>70$M users and ads campaigns demonstrate that CCWS achieves substantial improvements over several hashing-based methods including sign random projections (SignRP), minwise hashing (MinHash), as well as the vanilla CWS.

cs.IR↗

arXiv2022

Analysis of Error Feedback in Federated Non-Convex Optimization with Biased Compression

In federated learning (FL) systems, e.g., wireless networks, the communication cost between the clients and the central server can often be a bottleneck. To reduce the communication cost, the paradigm of communication compression has become a popular strategy in the literature. In this paper, we focus on biased gradient compression techniques in non-convex FL problems. In the classical setting of distributed learning, the method of error feedback (EF) is a common technique to remedy the downsides of biased gradient compression. In this work, we study a compressed FL scheme equipped with error feedback, named Fed-EF. We further propose two variants: Fed-EF-SGD and Fed-EF-AMS, depending on the choice of the global model optimizer. We provide a generic theoretical analysis, which shows that directly applying biased compression in FL leads to a non-vanishing bias in the convergence rate. The proposed Fed-EF is able to match the convergence rate of the full-precision FL counterparts under data heterogeneity with a linear speedup. Moreover, we develop a new analysis of the EF under partial client participation, which is an important scenario in FL. We prove that under partial participation, the convergence rate of Fed-EF exhibits an extra slow-down factor due to a so-called ``stale error compensation'' effect. A numerical study is conducted to justify the intuitive impact of stale error accumulation on the norm convergence of Fed-EF under partial participation. Finally, we also demonstrate that incorporating the two-way compression in Fed-EF does not change the convergence results. In summary, our work conducts a thorough analysis of the error feedback in federated non-convex optimization. Our analysis with partial client participation also provides insights on a theoretical limitation of the error feedback mechanism, and possible directions for improvements.

stat.ML↗

arXiv2022

$k$-Median Clustering via Metric Embedding: Towards Better Initialization with Differential Privacy

When designing clustering algorithms, the choice of initial centers is crucial for the quality of the learned clusters. In this paper, we develop a new initialization scheme, called HST initialization, for the $k$-median problem in the general metric space (e.g., discrete space induced by graphs), based on the construction of metric embedding tree structure of the data. From the tree, we propose a novel and efficient search algorithm, for good initial centers that can be used subsequently for the local search algorithm. Our proposed HST initialization can produce initial centers achieving lower errors than those from another popular initialization method, $k$-median++, with comparable efficiency. The HST initialization can also be extended to the setting of differential privacy (DP) to generate private initial centers. We show that the error from applying DP local search followed by our private HST initialization improves previous results on the approximation error, and approaches the lower bound within a small factor. Experiments justify the theory and demonstrate the effectiveness of our proposed method. Our approach can also be extended to the $k$-means problem.

cs.DS↗

arXiv2022

Layer-wise and Dimension-wise Locally Adaptive Federated Learning

In the emerging paradigm of Federated Learning (FL), large amount of clients such as mobile devices are used to train possibly high-dimensional models on their respective data. Combining (dimension-wise) adaptive gradient methods (e.g. Adam, AMSGrad) with FL has been an active direction, which is shown to outperform traditional SGD based FL in many cases. In this paper, we focus on the problem of training federated deep neural networks, and propose a novel FL framework which further introduces layer-wise adaptivity to the local model updates. Our framework can be applied to locally adaptive FL methods including two recent algorithms, Mime and Fed-AMS. Theoretically, we provide a convergence analysis of our layer-wise FL methods, coined Fed-LAMB and Mime-LAMB, which matches the convergence rate of state-of-the-art results in FL and exhibits linear speedup in terms of the number of workers. Experimental results on various datasets and models, under both IID and non-IID local data settings, show that both Fed-LAMB and Mime-LAMB achieve faster convergence speed and better generalization performance, compared to the various recent adaptive FL methods.

cs.LG↗

arXiv2022

Asymptotic Independence of the Sum and Maximum of Dependent Random Variables with Applications to High-Dimensional Tests

For a set of dependent random variables, without stationary or the strong mixing assumptions, we derive the asymptotic independence between their sums and maxima. Then we apply this result to high-dimensional testing problems, where we combine the sum-type and max-type tests and propose a novel test procedure for the one-sample mean test, the two-sample mean test and the regression coefficient test in high-dimensional setting. Based on the asymptotic independence between sums and maxima, the asymptotic distributions of test statistics are established. Simulation studies show that our proposed tests have good performance regardless of data being sparse or not. Examples on real data are also presented to demonstrate the advantages of our proposed methods.

stat.ME↗

arXiv2022

On Distributed Adaptive Optimization with Gradient Compression

We study COMP-AMS, a distributed optimization framework based on gradient averaging and adaptive AMSGrad algorithm. Gradient compression with error feedback is applied to reduce the communication cost in the gradient transmission process. Our convergence analysis of COMP-AMS shows that such compressed gradient averaging strategy yields same convergence rate as standard AMSGrad, and also exhibits the linear speedup effect w.r.t. the number of local workers. Compared with recently proposed protocols on distributed adaptive methods, COMP-AMS is simple and convenient. Numerical experiments are conducted to justify the theoretical findings, and demonstrate that the proposed method can achieve same test accuracy as the full-gradient AMSGrad with substantial communication savings. With its simplicity and efficiency, COMP-AMS can serve as a useful distributed training framework for adaptive gradient methods.

stat.ML↗

arXiv2022

Breaking the Linear Error Barrier in Differentially Private Graph Distance Release

Releasing all pairwise shortest path (APSP) distances between vertices on general graphs under weight Differential Privacy (DP) is known as a challenging task. In the previous attempt of (Sealfon 2016}, by adding Laplace noise to each edge weight or to each output distance, to achieve DP with some fixed budget, with high probability the maximal absolute error among all published pairwise distances is roughly $O(n)$ where $n$ is the number of nodes. It was shown that this error could be reduced for some special graphs, which, however, is hard for general graphs. Therefore, whether the approximation error can be reduced to sublinear in $n$ is posted as an interesting open problem. We break the linear barrier on the distance approximation error of previous result, by proposing an algorithm that releases a constructed synthetic graph privately. Computing all pairwise distances on the constructed graph only introduces $\tilde O(n^{1/2})$ error in answering all pairwise shortest path distances for fixed privacy parameter. Our method is based on a novel graph diameter (link length) augmentation via constructing "shortcuts" for the paths. By adding a set of shortcut edges to the original graph, we show that any node pair has a shortest path with link length $\tilde O(n^{1/2})$. Then by adding noises with some positive mean to the edge weights, we show that the new graph is differentially private and can be published to answer all pairwise shortest path distances with $\tilde O(n^{1/2})$ approximation error using standard APSP computation. Additionally, we consider the graph with small feedback vertex set number. A feedback vertex set (FVS) of a graph is a set of vertices whose removal leaves a graph without cycles, and the feedback vertex set number of a graph, $k$, is the size of a smallest feedback vertex set. We propose a DP algorithm with error rate $\tilde O(k)$.

cs.DS↗

arXiv2022

Communication-Efficient TeraByte-Scale Model Training Framework for Online Advertising

Click-Through Rate (CTR) prediction is a crucial component in the online advertising industry. In order to produce a personalized CTR prediction, an industry-level CTR prediction model commonly takes a high-dimensional (e.g., 100 or 1000 billions of features) sparse vector (that is encoded from query keywords, user portraits, etc.) as input. As a result, the model requires Terabyte scale parameters to embed the high-dimensional input. Hierarchical distributed GPU parameter server has been proposed to enable GPU with limited memory to train the massive network by leveraging CPU main memory and SSDs as secondary storage. We identify two major challenges in the existing GPU training framework for massive-scale ad models and propose a collection of optimizations to tackle these challenges: (a) the GPU, CPU, SSD rapidly communicate with each other during the training. The connections between GPUs and CPUs are non-uniform due to the hardware topology. The data communication route should be optimized according to the hardware topology; (b) GPUs in different computing nodes frequently communicates to synchronize parameters. We are required to optimize the communications so that the distributed system can become scalable. In this paper, we propose a hardware-aware training workflow that couples the hardware topology into the algorithm design. To reduce the extensive communication between computing nodes, we introduce a $k$-step model merging algorithm for the popular Adam optimizer and provide its convergence rate in non-convex optimization. To the best of our knowledge, this is the first application of $k$-step adaptive optimization method in industrial-level CTR model training. The numerical results on real-world data confirm that the optimized system design considerably reduces the training time of the massive model, with essentially no loss in accuracy.

cs.IR↗

arXiv2021

C-OPH: Improving the Accuracy of One Permutation Hashing (OPH) with Circulant Permutations

Minwise hashing (MinHash) is a classical method for efficiently estimating the Jaccrad similarity in massive binary (0/1) data. To generate $K$ hash values for each data vector, the standard theory of MinHash requires $K$ independent permutations. Interestingly, the recent work on "circulant MinHash" (C-MinHash) has shown that merely two permutations are needed. The first permutation breaks the structure of the data and the second permutation is re-used $K$ time in a circulant manner. Surprisingly, the estimation accuracy of C-MinHash is proved to be strictly smaller than that of the original MinHash. The more recent work further demonstrates that practically only one permutation is needed. Note that C-MinHash is different from the well-known work on "One Permutation Hashing (OPH)" published in NIPS'12. OPH and its variants using different "densification" schemes are popular alternatives to the standard MinHash. The densification step is necessary in order to deal with empty bins which exist in One Permutation Hashing. In this paper, we propose to incorporate the essential ideas of C-MinHash to improve the accuracy of One Permutation Hashing. Basically, we develop a new densification method for OPH, which achieves the smallest estimation variance compared to all existing densification schemes for OPH. Our proposed method is named C-OPH (Circulant OPH). After the initial permutation (which breaks the existing structure of the data), C-OPH only needs a "shorter" permutation of length $D/K$ (instead of $D$), where $D$ is the original data dimension and $K$ is the total number of bins in OPH. This short permutation is re-used in $K$ bins in a circulant shifting manner. It can be shown that the estimation variance of the Jaccard similarity is strictly smaller than that of the existing (densified) OPH methods.

stat.ML↗

arXiv2021

C-MinHash: Practically Reducing Two Permutations to Just One

Traditional minwise hashing (MinHash) requires applying $K$ independent permutations to estimate the Jaccard similarity in massive binary (0/1) data, where $K$ can be (e.g.,) 1024 or even larger, depending on applications. The recent work on C-MinHash (Li and Li, 2021) has shown, with rigorous proofs, that only two permutations are needed. An initial permutation is applied to break whatever structures which might exist in the data, and a second permutation is re-used $K$ times to produce $K$ hashes, via a circulant shifting fashion. (Li and Li, 2021) has proved that, perhaps surprisingly, even though the $K$ hashes are correlated, the estimation variance is strictly smaller than the variance of the traditional MinHash. It has been demonstrated in (Li and Li, 2021) that the initial permutation in C-MinHash is indeed necessary. For the ease of theoretical analysis, they have used two independent permutations. In this paper, we show that one can actually simply use one permutation. That is, one single permutation is used for both the initial pre-processing step to break the structures in the data and the circulant hashing step to generate $K$ hashes. Although the theoretical analysis becomes very complicated, we are able to explicitly write down the expression for the expectation of the estimator. The new estimator is no longer unbiased but the bias is extremely small and has essentially no impact on the estimation accuracy (mean square errors). An extensive set of experiments are provided to verify our claim for using just one permutation.

cs.DS↗

arXiv2021

Toward Communication Efficient Adaptive Gradient Method

In recent years, distributed optimization is proven to be an effective approach to accelerate training of large scale machine learning models such as deep neural networks. With the increasing computation power of GPUs, the bottleneck of training speed in distributed training is gradually shifting from computation to communication. Meanwhile, in the hope of training machine learning models on mobile devices, a new distributed training paradigm called ``federated learning'' has become popular. The communication time in federated learning is especially important due to the low bandwidth of mobile devices. While various approaches to improve the communication efficiency have been proposed for federated learning, most of them are designed with SGD as the prototype training algorithm. While adaptive gradient methods have been proven effective for training neural nets, the study of adaptive gradient methods in federated learning is scarce. In this paper, we propose an adaptive gradient method that can guarantee both the convergence and the communication efficiency for federated learning.

cs.LG↗