亚洲男人的天堂2018av,欧美草比,久久久久久免费视频精选,国色天香在线看免费,久久久久亚洲av成人片仓井空

Several applications in time series forecasting require predicting multiple steps ahead. Despite the vast amount of literature in the topic, both classical and recent deep learning based approaches have mostly focused on minimising performance averaged over the predicted window. We observe that this can lead to disparate distributions of errors across forecasting steps, especially for recent transformer architectures trained on popular forecasting benchmarks. That is, optimising performance on average can lead to undesirably large errors at specific time-steps. In this work, we present a Constrained Learning approach for long-term time series forecasting that aims to find the best model in terms of average performance that respects a user-defined upper bound on the loss at each time-step. We call our approach loss shaping constraints because it imposes constraints on the loss at each time step, and leverage recent duality results to show that despite its non-convexity, the resulting problem has a bounded duality gap. We propose a practical Primal-Dual algorithm to tackle it, and demonstrate that the proposed approach exhibits competitive average performance in time series forecasting benchmarks, while shaping the distribution of errors across the predicted window.

Mega-constellations of small satellites have evolved into a source of massive amount of valuable data. To manage this data efficiently, on-board federated learning (FL) enables satellites to train a machine learning (ML) model collaboratively without having to share the raw data. This paper introduces a scheme for scheduling on-board FL for constellations connected with intra-orbit inter-satellite links. The proposed scheme utilizes the predictable visibility pattern between satellites and ground station (GS), both at the individual satellite level and cumulatively within the entire orbit, to mitigate intermittent connectivity and best use of available time. To this end, two distinct schedulers are employed: one for coordinating the FL procedures among orbits, and the other for controlling those within each orbit. These two schedulers cooperatively determine the appropriate time to perform global updates in GS and then allocate suitable duration to satellites within each orbit for local training, proportional to usable time until next global update. This scheme leads to improved test accuracy within a shorter time.

Extremely large-scale massive multiple-input multiple-output (XL-MIMO) systems introduce the much higher channel dimensionality and incur the additional near-field propagation effect, aggravating the computation load and the difficulty to acquire the prior knowledge for channel estimation. In this article, an XL-MIMO channel network (XLCNet) is developed to estimate the high-dimensional channel, which is a universal solution for both the near-field users and far-field users with different channel statistics. Furthermore, a compressed XLCNet (C-XLCNet) is designed via weight pruning and quantization to accelerate the model inference as well as to facilitate the model storage and transmission. Simulation results show the performance superiority and universality of XLCNet. Compared to XLCNet, C-XLCNet incurs the limited performance loss while reducing the computational complexity and model size by about $10 \times$ and $36 \times$, respectively.

Reduced-rank regression recognises the possibility of a rank-deficient matrix of coefficients. We propose a novel Bayesian model for estimating the rank of the coefficient matrix, which obviates the need for post-processing steps and allows for uncertainty quantification. Our method employs a mixture prior on the regression coefficient matrix along with a global-local shrinkage prior on its low-rank decomposition. Then, we rely on the Signal Adaptive Variable Selector to perform sparsification and define two novel tools: the Posterior Inclusion Probability uncertainty index and the Relevance Index. The validity of the method is assessed in a simulation study, and then its advantages and usefulness are shown in real-data applications on the chemical composition of tobacco and on the photometry of galaxies.

Transformers generalize to novel compositions of structures and entities after being trained on a complex dataset, but easily overfit on datasets of insufficient complexity. We observe that when the training set is sufficiently complex, the model encodes sentences that have a common syntactic structure using a systematic attention pattern. Inspired by this observation, we propose SQ-Transformer (Structurally Quantized) that explicitly encourages systematicity in the embeddings and attention layers, even with a training set of low complexity. At the embedding level, we introduce Structure-oriented Vector Quantization (SoVQ) to cluster word embeddings into several classes of structurally equivalent entities. At the attention level, we devise the Systematic Attention Layer (SAL) and an alternative, Systematically Regularized Layer (SRL) that operate on the quantized word embeddings so that sentences of the same structure are encoded with invariant or similar attention patterns. Empirically, we show that SQ-Transformer achieves stronger compositional generalization than the vanilla Transformer on multiple low-complexity semantic parsing and machine translation datasets. In our analysis, we show that SoVQ indeed learns a syntactically clustered embedding space and SAL/SRL induces generalizable attention patterns, which lead to improved systematicity.

Probabilistic forecasting of irregularly sampled multivariate time series with missing values is an important problem in many fields, including health care, astronomy, and climate. State-of-the-art methods for the task estimate only marginal distributions of observations in single channels and at single timepoints, assuming a fixed-shape parametric distribution. In this work, we propose a novel model, ProFITi, for probabilistic forecasting of irregularly sampled time series with missing values using conditional normalizing flows. The model learns joint distributions over the future values of the time series conditioned on past observations and queried channels and times, without assuming any fixed shape of the underlying distribution. As model components, we introduce a novel invertible triangular attention layer and an invertible non-linear activation function on and onto the whole real line. We conduct extensive experiments on four datasets and demonstrate that the proposed model provides $4$ times higher likelihood over the previously best model.

Language models (LMs) have already demonstrated remarkable abilities in understanding and generating both natural and formal language. Despite these advances, their integration with real-world environments such as large-scale knowledge bases (KBs) remains an underdeveloped area, affecting applications such as semantic parsing and indulging in "hallucinated" information. This paper is an experimental investigation aimed at uncovering the robustness challenges that LMs encounter when tasked with knowledge base question answering (KBQA). The investigation covers scenarios with inconsistent data distribution between training and inference, such as generalization to unseen domains, adaptation to various language variations, and transferability across different datasets. Our comprehensive experiments reveal that even when employed with our proposed data augmentation techniques, advanced small and large language models exhibit poor performance in various dimensions. While the LM is a promising technology, the robustness of the current form in dealing with complex environments is fragile and of limited practicality because of the data distribution issue. This calls for future research on data collection and LM learning paradims.

Reconfigurable intelligent surface (RIS) is a promising solution to deal with the blockage-sensitivity of millimeter wave band and reduce the high energy consumption caused by network densification. However, deploying large scale RISs may not bring expected performance gain due to significant channel estimation overhead and non-negligible reflected interference. In this paper, we derive the analytical expressions of the coverage probability, area spectrum efficiency (ASE) and energy efficiency (EE) of a downlink RIS-aided multi-cell network. In order to optimize the network performance, we investigate the conditions for the optimal number of training symbols of each antenna-to-antenna and antenna-to-element path (referred to as the optimal unit training overhead) in channel estimation. Our study shows that: 1) RIS deployment is not `the more, the better', only when blockage objects are dense should one deploy more RISs; 2) the coverage probability is maximized when the unit training overhead is designed as large as possible; 3) however, the ASE-and-EE-optimal unit training overhead exists. It is a monotonically increasing function of the frame length and a monotonically decreasing function of the average signal-to-noise-ratio (in the high signal-to-noise-ratio region). Additionally, the optimal unit training overhead is smaller when communication ends deploy particularly few or many antennas.

The paper aims to address load imbalance caused by high in-degree distribution in graphs by applying the idea of rhizome to vertex-centric message-driven graph processing. Rhizome construction of the graph creates multiple named vertex address for any number of single large in-degree vertices. It then allows other vertices to point to any of the named addresses thus sharing the in-degree load. The rhizomes internally communicate and remain consistent to provide a unified and correct view of the vertex. Simulated experimental results show performance speed ups for BFS graph traversal on large chip sizes for the tested input graph datasets containing highly skewed in-degree distribution. The improvements come from sharing the in-degree compute workload among memory-processing elements and also lowering contention on the network-on-chip.

Message brokers often mediate communication between data producers and consumers by adding variable-sized messages to ordered distributed queues. Our goal is to determine the number of consumers and consumer-partition assignments needed to ensure that the rate of data consumption keeps up with the rate of data production. We model the problem as a variable item size bin packing problem. As the rate of production varies, new consumer-partition assignments are computed, which may require rebalancing a partition from one consumer to another. While rebalancing a queue, the data being produced into the queue is not read leading to additional latency costs. As such, we focus on the multi-objective optimization cost of minimizing both the number of consumers and queue migrations. We present a variety of algorithms and compare them to established bin packing heuristics for this application. Comparing our proposed consumer group assignment strategy with Kafka's, a commonly employed strategy, our strategy presents a 90th percentile latency of 4.52s compared to Kafka's 217s with both using the same amount of consumers. Kafka's assignment strategy only improved the consumer group's performance with regards to latency with configurations that used at least 60% more resources than our approach.