A connectional brain template (CBT) is a holistic representation of a population of brain connectivities. The federated learning of CBT allows for estimating the CBT of brain connectivities from multiple domains (i.e., hospitals) in a fully data-preserving manner. However, existing methods overlook the non-independent and identically distributed (non-IID) issue stemming from the heterogeneity of multi-domain brain connectivities. This non-IID issue degrades the centrality of locally learned CBT from multiple decentralized domains, eventually leading to a limited representation ability. To overcome this limitation, we propose a metadata-driven federated learning framework, called MetaFedCBT, for multi-domain CBT learning under the non-IID condition. Given the data drawn from a specific domain, our model is able to predict the metadata (i.e., statistics) of other unseen domains with a proposed metadata regressor and local-global network residual weights. Furthermore, we introduce a metadata-driven connectivity generator to predict brain connectivities of unseen domains under the guidance of obtained metadata. As the federated learning progresses over multiple rounds, we continuously update the predicted metadata and brain connectivities to better approximate the unseen domains. MetaFedCBT overcomes the non-IID issue by generating informative brain connectivities for privacy-preserving holistic CBT learning. Extensive experiments on multi-view morphological brain networks of normal and patient subjects demonstrate that our MetaFedCBT is a superior federated CBT learning model and significantly advances state-of-the-art performance.

In modern datacenters, the effectiveness of end-to-end congestion control (CC) is quickly diminishing with the rapid bandwidth evolution. Per-hop flow control (FC) can react to congestion more promptly. However, a coarse-grained FC can result in Head-Of-Line (HOL) blocking. A fine-grained, per-flow FC can eliminate HOL blocking caused by flow control, however, it does not scale well. This paper presents Pyrrha, a scalable flow control approach that provably eliminates HOL blocking while using a minimum number of queues. In Pyrrha, flow control first takes effect on the root of the congestion, i.e., the port where congestion occurs. And then flows are controlled according to their contributed congestion roots. A prototype of Pyrrha is implemented on Tofino2 switches. Compared with state-of-the-art approaches, the average FCT of uncongested flows is reduced by 42%-98%, and 99th-tail latency can be$1.6\times $-$215\times $lower, without compromising the performance of congested flows.

Surface magnetic resonance sounding (MRS) offers the advantages of direct, quantitative, and unique interpretations in the field of groundwater detection. The time-lapse inversion (TLI) method, with its temporal continuity, has been applied to monitor the time-varying trends of the hydrological parameters of groundwater. However, the ambient noise levels in MRS data fluctuate significantly over time (daily), and the presence of low signal-to-noise ratio (SNR) data can lead to a deterioration in the results of TLI. Thus, we propose a new TLI of MRS data weighted by noise-level estimation in this article. Noise-weighted TLI (NW-TLI) quantifies the reliability of each MRS dataset on the basis of noise estimation residuals and incorporates time-lapse reference weights into the inversion process, thereby ensuring that the hydrological trends are more reasonably constrained by high-SNR data. In synthetic data experiments, we demonstrate that the NW-TLI method effectively mitigates interference from adjacent low-SNR data under various complicated noisy cases. Even with multiple sets of low-SNR MRS data, NW-TLI can provide more accurate hydrological time-varying trends than conventional TLI. Additionally, we assess the impact of the temporal variability of the water-bearing model and the degree of data weighting on the interpretative accuracy and ultimately validate the practicability of the NW-TLI method via field-measured data.

RGB-D salient object detection (SOD) has gained tremendous attention in recent years. In particular, transformer has been employed and shown great potential. However, existing transformer models usually overlook the vital edge information, which is a major issue restricting the further improvement of SOD accuracy. To this end, we propose a novel edge-aware RGB-D SOD transformer, called EM-Trans, which explicitly models the edge information in a dual-band decomposition framework. Specifically, we employ two parallel decoder networks to learn the high-frequency edge and low-frequency body features from the low- and high-level features extracted from a two-steam multimodal backbone network, respectively. Next, we propose a cross-attention complementarity exploration module to enrich the edge/body features by exploiting the multimodal complementarity information. The refined features are then fed into our proposed color-hint guided fusion module for enhancing the depth feature and fusing the multimodal features. Finally, the resulting features are fused using our deeply supervised progressive fusion module, which progressively integrates edge and body features for predicting saliency maps. Our model explicitly considers the edge information for accurate RGB-D SOD, overcoming the limitations of existing methods and effectively improving the performance. Extensive experiments on benchmark datasets demonstrate that EM-Trans is an effective RGB-D SOD framework that outperforms the current state-of-the-art models, both quantitatively and qualitatively. A further extension to RGB-T SOD demonstrates the promising potential of our model in various kinds of multimodal SOD tasks.

Congestion control protocols play a vital role in enhancing the performance of various applications within datacenter networks. While reactive congestion control (RCC) protocols are widely deployed in commercial datacenters, the research community has actively explored token-based proactive congestion control (TCC) protocols to further push the boundaries of performance. However, despite the emergence of numerous TCC variants, there has been a lack of systematic exploration in the design space of TCC. This paper aims to bridge this gap by proposing a framework for understanding the design choices within the TCC approach. In this study, we systematically analyze different design choices of TCC approaches and leverage this understanding to develop a novel TCC protocol called ToCC. To implement ToCC, we address a set of challenges and deploy it in NP-based smart NICs. We compare ToCC with state-of-the-art TCC and RCC protocols through extensive large-scale simulations and testbed evaluations. The results demonstrate that ToCC exhibits robustness in achieving low latency across various scenarios. Additionally, ToCC effectively reduces buffer occupancy by 4.8 times compared to existing approaches, and under incast scenarios, it significantly shortens flow completion time by up to 90%. Congestion control protocols are crucial for optimizing the performance of datacenter network applications. Although reactive congestion control (RCC) protocols are commonly used in commercial datacenters, researchers have been exploring token-based proactive congestion control (TCC) protocols to further enhance network performance. Despite the development of numerous TCC variants, there has not been a thorough examination of the design space of TCC protocols until now. This paper aims to address this gap by introducing a framework for understanding the design choices within the TCC approach for TCC protocols. By analyzing various design aspects of TCC approaches, we create a novel TCC protocol called ToCC. At the central of ToCC design is that it leverages congestion control mechanisms over tokens. To implement ToCC, we tackle several challenges and integrate it into NP-based smart NICs. Comparing ToCC with state-of-the-art TCC and RCC protocols through extensive large-scale simulations and testbed evaluations, we find that ToCC consistently achieves low latency across different scenarios. Moreover, ToCC significantly reduces buffer occupancy by 4.8 times compared to existing methods, and during incast scenarios, it decreases flow completion time by up to 90%.

Yanhe Fu, Yanan Cao, Qingyue Wang, Yi Liu. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.

Answering complex logical queries over knowledge graphs (KGs) is a fundamental yet challenging task. Recently, query representation has been a mainstream approach to complex logical reasoning, making the target answer and query closer in the embedding space. However, there are still two limitations. First, prior methods model the query as a fixed vector, but ignore the uncertainty of relations on KGs. In fact, different relations may contain different semantic distributions. Second, traditional representation frameworks fail to capture the joint distribution of queries and answers, which can be learned by generative models that have the potential to produce more coherent answers. To alleviate these limitations, we propose a novel generative model, named DiffCLR, which exploits the diffusion model for complex logical reasoning to approximate query distributions. Specifically, we first devise a query transformation to convert logical queries into input sequences by dynamically constructing contextual subgraphs. Then, we integrate them into the diffusion model to execute a multi-step generative process, and a structure-enhanced self-attention is further designed for incorporating the structural features embodied in KGs. Experimental results on two benchmark datasets show our model effectively outperforms state-of-the-art methods, particularly in multi-hop chain queries with significant improvement.

Magnetic resonance tomography (MRT) is a geophysical exploration technique that enables the imaging of 2-D or 3-D water-bearing structures, offering distinct advantages, including noninvasiveness, quantifiability, and unique interpretability. Currently, MRT data inversion mainly relies on the Q-time (QT) inversion method. Since this method utilizes the Gauss-Newton iteration to seek the optimal solution, it involves a considerable amount of computational workload, thus consuming a significant amount of time. To overcome this challenge, this study introduces an accelerated imaging method by combining the singular value decomposition (SVD) pseudoinversion algorithm and the deep neural network algorithm. The SVD pseudoinversion algorithm transforms MRT data into a water-bearing feature matrix containing only water content and relaxation time information by introducing a priori forward kernel function. Subsequently, neural network establishes a nonlinear mapping relationship between the water-bearing feature matrix and the spatial distribution of the water content and relaxation time in the subsurface. The SVD pseudoinversion algorithm, by incorporating prior information, mitigates the distribution differences in MRT data caused by geological and measurement parameters. This addresses the limited applicability of deep learning methods under complex geological conditions and multiple measurement schemes. The experimental results demonstrate that the method achieves precise and rapid imaging, while also possessing effectiveness and practicality.

Qingyue Wang, Liang Ding, Yanan Cao, Yibing Zhan, Zheng Lin, Shi Wang, Dacheng Tao, Li Guo. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

The average/tail Flow Completion Times (FCTs) are critical to many datacenter applications. Congestion control plays a central role in optimizing FCT. Inappropriate congestion control can exacerbate buffer occupancy, thus hurting the flow performance. Our observations are that current approaches are too aggressive in injecting packets into underlying networks. Instead of handling buffer explosion afterward, we reduce buffer occupancy in the first place. We propose PayDebt, a novel and readily-deployable proactive congestion control protocol. At its heart, adebtmechanism provides bandwidth coordination between the already-buffered and the forthcoming packets. We evaluate PayDebt both in a testbed and large-scale simulations. The buffer occupancy can be decreased by up to 8.0×-35.9× compared to DCQCN and Homa.

Incast occurs frequently in datacenter networks where a large number of senders send data to a single receiver simultaneously, which makes the last hop the network bottleneck. Incast can hurt flows' performance. However, congestion control protocols are not effective at handling incast. One key insight is that it is too late to handle incast packets after they have already piled up at the last hop. Instead, we should avoid incast as early as possible. Inspired by flood control in Hydrologic Engineering, we propose Floodgate, a novel switch-based per-hop flow control to handle incast. Floodgate is compatible with existing congestion control protocols. We integrate it with practical congestion control approaches such as DCQCN, TIMELY, and HPCC. We evaluate Floodgate both in our implementations and large-scale simulations. Compared with state of the art, Floodgate reduces the buffer occupancy by a factor of 6.6x, as well as the queuing delay. Therefore, the average FCT and tail latency are greatly reduced.

External knowledge is vital to many natural language processing tasks. However, current end-to-end dialogue systems often struggle to interface knowledge bases(KBs) with response smoothly and effectively. In this paper, we convert the raw knowledge into relation knowledge and integrated knowledge and then incorporate them into end-to-end task-oriented dialogue systems. The relation knowledge extracted from knowledge triples is combined with dialogue history, aiming to enhance semantic inputs and support better language understanding. Integrated knowledge involves entities and relations by graph attention, assisting the model in generating informative responses. The experimental results on three public dialogue datasets show that our model improves over the previous state-of-the-art models in sentence fluency and informativeness.

In memory computing and high-end distributed storage demand low latency, high throughput, and zero data loss simultaneously from datacenter networks. Existing reactive congestion control approaches cannot both minimize queuing latency and ensure zero data loss. A token-oriented proactive approach can achieve them together by controlling congestion even before sending data packets. However, state-of-the-art token-oriented approaches only strive to optimize network-level metrics: maximizing throughput while achieving flow-level fairness. This article answers the question of how to support objective-aware traffic scheduling in token-oriented approaches. The novelty of Token-Oriented in-network Prioritization (TOP) is that it prioritizes tokens instead of data packets. We make three contributions. Via simulations over a hypothetical TOP system, our first contribution is demonstrating the potential performance gain that can be brought by TOP. Second, we investigate the applicability of TOP. Although the overhead of enabling necessary TOP features in switches is trivial, we find that mainstream commodity datacenter switches do not support them. We hence propose a readily-deployable remedy to achieve in-network prioritization by pushing both switch and end-host hardware capacity to an extreme end. Lastly, we implement a running TOP system with Linux hosts and commodity switches, and evaluate TOP in testbeds and with large-scale simulations for various scenarios.