The surge in large language models (LLMs) has fundamentally reshaped the landscape of GPU usage patterns, creating an urgent need for more efficient management strategies. While cloud providers employ spot instances to reduce costs for low-priority (LP) tasks, existing schedulers still grapple with high eviction rates and lengthy queuing times. To address these limitations, we present GFS, a novel preemptive scheduling framework that enhances service-level objective (SLO) compliance for high-priority (HP) tasks while minimizing preemptions to LP tasks. Firstly, GFS utilizes a lightweight forecasting model that predicts GPU demand among different tenants, enabling proactive resource management. Secondly, GFS employs a dynamic allocation mechanism to adjust the spot quota for LP tasks with guaranteed durations. Lastly, GFS incorporates a preemptive scheduling policy that prioritizes HP tasks while minimizing the impact on LP tasks. We demonstrate the effectiveness of GFS through both real-world implementation and simulations. The results show that GFS reduces eviction rates by 33.0%, and cuts queuing delays by 44.1% for LP tasks. Furthermore, GFS enhances the GPU allocation rate by up to 22.8% in real production clusters. In a production cluster of more than 10,000 GPUs, GFS yields roughly $459,715 in monthly benefits.

Practical applications often require forecasting the future states of short time series (STS) using multiple related long time series (LTS) as auxiliary data, a process known as few-shot prediction. The primary challenge, given the limited data on STS, is effectively capturing the pattern similarities between STS and LTS. Current methods, despite notable advancements, primarily focus on trans- ferring pattern characteristics from LTS to STS without explicitly addressing their similarities at various levels. To overcome this lim- itation, we propose a novel few-shot time series forecasting model called DANet. Drawing on the Retrieval-Augmented Generation (RAG) framework in large language models, DANet retrieves long and short sequences from LTS that closely resemble STS, thereby enhancing prediction accuracy while simultaneously reducing un- certainty through this retrieval process. First, we define two metrics to quantify pattern similarities between STS and LTS, addressing the issue of different representations of the same pattern due to variations in sequence length. Second, we propose a dual-attention mechanism which embeds the two similarities metrics to extract and integrate long and short sequences from LTS across variable and temporal levels for generating predictions. Our experiments across six scenarios show that DANet significantly outperforms six state-of-the-art (SOTA) methods.

Flexible intelligent metasurface (FIM) technology has emerged as a promising solution for enhancing wireless communication performance. In contrast to traditional rigid reconfigurable intelligent surfaces (RIS), an FIM consists of an array of electromagnetic (EM) elements, each capable of flexibly adjusting its position along the direction perpendicular to the surface to collaboratively morph the surface shape. In this paper, an optimization problem for maximizing the weighted sum-rate (WSR) in an FIM-aided multicell multi-user multiple-input single-output (MU-MISO) system is investigated. We jointly optimize the beamforming at the base station (BS), the phase shift matrix, and the FIM surface shape. To address this problem, we propose an efficient alternating optimization framework, where we employ the weighted minimum mean square error (WMMSE) method to reformulate the problem and the block coordinate descent (BCD) algorithm to iteratively update the variables. Specifically, we utilize the Riemannian Conjugate Gradient (RCG) algorithm to optimize the phase shift matrix, and the projected gradient descent (PGD) method to optimize the FIM surface shape. Additionally, the optimal beamforming vectors are obtained in closed form. Finally, simulation results demonstrate the superiority of FIM over conventional RIS in various scenarios.

Reconfigurable intelligent surface (RIS) technology holds immense potential for increasing the performance of wireless networks. Therefore, RIS is also regarded as one of the solutions to address communication challenges in high-mobility scenarios, such as Doppler shift and fast fading. This paper investigates a high-speed train (HST) multiple-input single-output (MISO) communication system aided by a RIS. We propose a block coordinate descent (BCD) algorithm to jointly optimize the RIS phase shifts and the transmit beamforming vectors to maximize the channel gain. Numerical results are provided to demonstrate that the proposed algorithm significantly enhances the system performance, achieving an average channel gain improvement of 15 dB compared to traditional schemes. Additionally, the introduction of RIS eliminates outage probability and improves key performance metrics such as achievable rate, channel capacity, and bit error rate (BER). These findings highlight the critical role of RIS in enhancing HST communication systems.

Map matching is crucial to facilitating location-based services, and recent advancements in map matching have demonstrated excellent performance with high-quality data. However, the use of low-precision devices often introduces high measurement noise, and the slow update rate of maps may result in errors in digital maps. Consequently, multiple types of noise significantly impact the performance of map matching algorithms. To tackle this issue, this paper presents a novel multi-noise perception framework, named MPF, aiming to enhance the performance and robustness of existing map matching algorithms. The main challenge lies in detecting anomalies during map matching, identifying the root causes, and devising appropriate solutions. Firstly, we propose a matching quality assessment (MQA) method that assesses abnormal variance in matching probability. Secondly, we introduce a multiple noise discrimination (MND) mechanism to effectively differentiate between measurement noise and map errors. Thirdly, we present a missing segment generation (MSG) scheme that dynamically fills in map gaps to prevent significant detours. To validate the effectiveness of MPF, we conduct experiments using real-world taxi trajectories from four cities, covering a total distance of 79,670.6 km. MPF is compare with seven online map matching algorithms and is used to optimize their performance. The experiments show that MPF outperforms the top baselines by 15.6%-26.9% and enhances their performance by 18.7%-38.2%.

Multivariate time series forecasting is applied in many domains, such as finance, transportation, and industry. The main challenge of precise forecasting lies in accurately capturing latent dependencies. Recent studies develop various frameworks to reduce computational complexity or to enhance the learning of intricate relationships, while lacking interpretability and generality. In this article, we aim to elucidate the capture of dependencies as the recognition of patterns. We believe that patterns can be formally described from two aspects: the shapes of segments that frequently repeat and the corresponding forms of repetitions. Drawing upon this idea, we design a multivariate time series forecasting model named PRformer , 1 which incorporates a pattern-oriented attention mechanism and a pattern-based projector. The attention mechanism can perceive different forms of repetitions by embedded with various similarity evaluation metrics between segments, and filter out noise from segments to extract potential patterns with a statistical-driven weighting scheme. The pattern-based projector is employed to form the forecasting results by deriving the representative patterns from the set of potential ones. By incorporating explicit definitions of patterns, PRformer is interpretable and general to various time series scenarios. Experimental results on seven datasets demonstrate that PRformer outperforms six state-of-the-art models by about 10.7% in forecasting accuracy.

Data imputation is crucial in the analysis of incomplete time series, such as forecasting and classification, which involves learning dependencies among the observed values to infer missing ones. As there are no ground truths for missing values, the challenge of time series imputation lies in preventing the model from overfitting to spurious correlations. In this article, we believe that ensuring dependency consistency between observed and imputed values in a sequence is paramount for data imputation. Based on this idea, we propose a model called IR 2 -Net , 1 which combines an incomplete representation mechanism (IRM) with an iterative reconstruction framework (IRF) to establish a closed-loop learning-validation imputation paradigm. Firstly, IRM facilitates the representation of dependencies in incomplete sequences while preserving their distributions and semantics, effectively preventing the model from capturing spurious correlations. Secondly, IRF enables the model to reconstruct identical complete sequences separately based on imputed and observed values, ensuring that the dependencies of imputed values remain consistent with those of the observed ones. We conduct experiments on four datasets and compare IR 2 -Net with seven state-of-the-art imputation models. The experiment results show that IR 2 -Net outperforms all the baselines by 4.1%–23.4% in terms of accuracy. Moreover, IRF and IRM are two general modules that can be easily integrated into two existing models, significantly enhancing their performance by 18.3%–42.0%.

Video gaming now represents the largest category in the entertainment industry in terms of revenue. To expand their market share, game developers are creating more cross-platform games, which are compatible with various platforms, including PCs, consoles, and smartphones. However, creating such games poses challenges as developers encounter platform-specific issues that may only surface on one of the target platforms. Consequently, many ported games fail due to careless adaptation from one exclusive platform to another. This paper presents the first empirical study on cross-platform issues by analyzing game users’ reviews for video games on both PC and game console(s). Our findings reveal that platform-related issues occur more frequently on the PC side, particularly for games that are ported from consoles. To address this challenge, we develop machine learning-based approaches to automatically identify and categorize reviews discussing platform-related issues, achieving a reasonable classification performance with 79.73% to 90.06% accuracy. Our approach would help cross-platform game developers save considerable time when analyzing user reviews.

Accurately forecasting workloads in terms of throughput that is quantified as queries per second (QPS) is essential for microservices to elastically adjust their resource allocations. However, long-term QPS prediction is challenging in two aspects: 1) generality across various services with different temporal patterns, 2) characterization of intricate QPS sequences which are entangled by multiple components. In this paper, we propose a knowledge auto-embedding Informer network (KAE-Informer) for forecasting the long-term QPS sequences of microservices. By analyzing a large number of microservice traces, we discover that there are two main decomposable and predictable components in QPS sequences, namely global trend & dominant periodicity (TP) and low-frequency residual patterns with long-range dependencies. These two components are important for accurately forecasting long-term QPS. First, KAE-Informer embeds the knowledge of TP components through mathematical modeling. Second, KAE-Informer designs a convolution ProbSparse self-attention mechanism and a multi-layer event discrimination scheme to extract and embed the knowledge of local context awareness and event regression effect implied in residual components, respectively. We conduct experiments based on three real datasets including a QPS dataset collected from 40 microservices. The experiment results show that KAE-Informer achieves a reduction of MAPE, MAE and RMSE by about 16.6%, 17.6% and 23.1% respectively, compared to the state-of-the-art models.

Online map matching is essential for some location-based services, such as car navigation. However, due to GPS measurement errors and/or the lack of sufficient information, the performance of most existing online algorithms will degrade in the increasingly complex traffic environment. In this paper, we propose an adaptive online map matching algorithm called AMM. The basic idea is that AMM should be able to calibrate GPS observation data for various measurement errors and under complex urban conditions. First, we establish a collaborative evaluation model between GPS points and candidate points to effectively filter low-quality GPS measurement points, dynamically set the weights of different features and comprehensively select the best candidate points. Second, we propose a retrospective correction mechanism to correct the previous matching results when more information is available, which will help improve the accuracy of future GPS points. Furthermore, we define parameter self-tuning rules for AMM to enhance its portability by avoiding time-consuming parameter tuning steps. We conduct extensive experiments to evaluate the performance of AMM on real vehicle trajectory datasets. The experiment results show that AMM outperforms its counterparts by up to 32% in terms of accuracy and its performance in different traffic conditions is more stable.

Data analysis reveals that passengers can only endure a few minutes before taking a taxi. However, most existing route recommendation systems are not adequate to satisfy impatient customers due to two shortcomings: inaccurate demand forecast and the lack of an efficient supply-demand balance mechanism. In this paper, we propose a recommendation system called @ Me, which aims to dispatch vacant taxis to the vicinity of potential customers at the right time. To achieve minute-level demand forecasting, we ensemble a contextualized spatial-temporal network (CSTN) with an LSTM network to optimize prediction accuracy. In addition, we characterize the attractiveness of the road grid to vacant taxis as a force model, on which a taxi scheduling algorithm is proposed to dynamically balance supply and demand. Extensive experiments on real datasets clearly indicate that our method is superior to the selected baselines. Vacant taxis that follow the routes suggested by @ME can catch more impatient customers in a shorter cruising time. The 7 -day experimental results on the Manhattan dataset show that @Me can carry an additional 48,332 passengers and increase drivers' revenue by $570,317.

With the growing market of cloud databases, careful detection and elimination of slow queries are of great importance to service stability. Previous studies focus on optimizing the slow queries that result from internal reasons (e.g., poorly-written SQLs). In this work, we discover a different set of slow queries which might be more hazardous to database users than other slow queries. We name such queries Intermittent Slow Queries (iSQs), because they usually result from intermittent performance issues that are external (e.g., at database or machine levels). Diagnosing root causes of iSQs is a tough but very valuable task. This paper presents iSQUAD, Intermittent Slow QUery Anomaly Diagnoser, a framework that can diagnose the root causes of iSQs with a loose requirement for human intervention. Due to the complexity of this issue, a machine learning approach comes to light naturally to draw the interconnection between iSQs and root causes, but it faces challenges in terms of versatility, labeling overhead and interpretability. To tackle these challenges, we design four components, i.e., Anomaly Extraction, Dependency Cleansing, Type-Oriented Pattern Integration Clustering (TOPIC) and Bayesian Case Model. iSQUAD consists of an offline clustering & explanation stage and an online root cause diagnosis & update stage. DBAs need to label each iSQ cluster only once at the offline stage unless a new type of iSQs emerges at the online stage. Our evaluations on real-world datasets from Alibaba OLTP Database show that iSQUAD achieves an iSQ root cause diagnosis average F1-score of 80.4%, and outperforms existing diagnostic tools in terms of accuracy and efficiency.

Oceans are important for scientific research and also for global economic and military security. Usually, wireless ad hoc networks are chosen to transform real-time data collected by ocean monitoring sensors (nodes). Due to the random motion of waves or the random direction of the wind, nodes in the network might become detached from the coverage of the network. In this case, the detached nodes can either send the collected data directly to the base station at the cost of consuming more energy or wait for a period of time to rejoin the network with the price of sacrificing the real time of the collected data. In this paper, we model the optimal waiting time for detached nodes before directly sending the data in the dynamic environment of ocean monitoring. For this purpose, we need to address two problems. The first is how to calculate the rate of coverage with a different number and different broadcast radii of nodes. The second is when a node detaches from the coverage of the network, how much time will it need to wait before it rejoins the network. We first establish the motion model of nodes, which is the basis to deduce the probability distribution of a certain time when the detached node rejoins the network. Based on the probability distribution, the waiting time of the detached nodes can be optimally determined, aiming to achieve a good balance between energy consumption and data timeliness. Finally, a series of simulations is conducted to validate the effectiveness of our proposed method.