Search

Temporal graph neural networks (TGNNs) outperform regular GNNs by incorporating time information into graph-based operations. However, TGNNs adopt specialized models (e.g., TGN, TGAT, and APAN ) and require tailored training frameworks (e.g., TGL and ETC). In this paper, we propose TF-TGN, which uses Transformer decoder as the backbone model for TGNN to enjoy Transformer's codebase for efficient training. In particular, Transformer achieves tremendous success for language modeling, and thus the community developed high-performance kernels (e.g., flash-attention and memory-efficient attention) and efficient distributed training schemes (e.g., PyTorch FSDP, DeepSpeed, and Megatron-LM). We observe that TGNN resembles language modeling, i.e., the message aggregation operation between chronologically occurring nodes and their temporal neighbors in TGNNs can be structured as sequence modeling. Beside this similarity, we also incorporate a series of algorithm designs including suffix infilling, temporal graph attention with self-loop, and causal masking self-attention to make TF-TGN work. During training, existing systems are slow in transforming the graph topology and conducting graph sampling. As such, we propose methods to parallelize the CSR format conversion and graph sampling. We also adapt Transformer codebase to train TF-TGN efficiently with multiple GPUs. We experiment with 9 graphs and compare with 2 state-of-the-art TGNN training frameworks. The results show that TF-TGN can accelerate training by over 2.20 while providing comparable or even superior accuracy to existing SOTA TGNNs. TF-TGN is available at https://github.com/qianghuangwhu/TF-TGN., Comment: conference Under review

Text-attributed graph (TAG) is an important type of graph structured data with text descriptions for each node. Few- and zero-shot node classification on TAGs have many applications in fields such as academia and social networks. However, the two tasks are challenging due to the lack of supervision signals, and existing methods only use the contrastive loss to align graph-based node embedding and language-based text embedding. In this paper, we propose Hound to improve accuracy by introducing more supervision signals, and the core idea is to go beyond the node-text pairs that come with data. Specifically, we design three augmentation techniques, i.e., node perturbation, text matching, and semantics negation to provide more reference nodes for each text and vice versa. Node perturbation adds/drops edges to produce diversified node embeddings that can be matched with a text. Text matching retrieves texts with similar embeddings to match with a node. Semantics negation uses a negative prompt to construct a negative text with the opposite semantics, which is contrasted with the original node and text. We evaluate Hound on 5 datasets and compare with 13 state-of-the-art baselines. The results show that Hound consistently outperforms all baselines, and its accuracy improvements over the best-performing baseline are usually over 5%.

Existing blockchain networks are often large-scale, requiring transactions to be synchronized across the entire network to reach consensus. On-chain computations can be prohibitively expensive, making many CPU-intensive computations infeasible. Inspired by the structure of IBM's token ring networks, we propose a lightweight consensus protocol called Monadring to address these issues. Monadring allows nodes within a large blockchain network to form smaller subnetworks, enabling faster and more cost-effective computations while maintaining the security guarantees of the main blockchain network. To further enhance Monadring's security, we introduce a node rotation mechanism based on Verifiable Random Function (VRF) and blind voting using Fully Homomorphic Encryption (FHE) within the smaller subnetwork. Unlike the common voting-based election of validator nodes, Monadring leverages FHE to conceal voting information, eliminating the advantage of the last mover in the voting process. This paper details the design and implementation of the Monadring protocol and evaluates its performance and feasibility through simulation experiments. Our research contributes to enhancing the practical utility of blockchain technology in large-scale application scenarios., Comment: 23 pages, 3 figures

In the past two decades, the exploration of quantumness within Schwarzschild spacetime has garnered significant interest, particularly regarding the Hawking radiation's impact on quantum correlations and quantum coherence. Building on this foundation, we investigate how Hawking radiation influences maximal steered coherence (MSC)-a crucial measure for gauging the ability to generate coherence through steering. We find that as the Hawking temperature increases, the physically accessible MSC degrade while the unaccessible MSC increase. This observation is attributed to a redistribution of the initial quantum correlations, previously acknowledged by inertial observers, across all bipartite modes. In particular, we find that in limit case that the Hawking temperature tends to infinity, the accessible MSC equals to 1/\sqrt{2} of its initial value, and the unaccessible MSC also equals to the same value. Our findings illuminate the intricate dynamics of quantum information in the vicinity of black holes, suggesting that Hawking radiation plays a pivotal role in reshaping the landscape of quantum coherence and entanglement in curved spacetime. This study not only advances our theoretical understanding of black hole thermodynamics but also opens new avenues for investigating the interface between quantum mechanics and general relativity., Comment: 4 pages, 1 figure

Superconductivity and magnetism are antagonistic quantum matter, while their intertwining has long been considered in frustrated-lattice systems1-3. In this work, we utilize scanning tunneling microscopy and muon spin resonance to discover time-reversal symmetry-breaking superconductivity in kagome metal Cs(V,Ta)3Sb5, where the Cooper pairing exhibits magnetism and is modulated by it. In the magnetic channel, we observe spontaneous internal magnetism in a full-gap superconducting state. Under perturbations of inverse magnetic fields, we detect a time-reversal asymmetrical interference of Bogoliubov quasi-particles at a circular vector. At this vector, the pairing gap spontaneously modulates, which is distinct from pair density waves occurring at a point vector and consistent with the theoretical proposal of unusual interference effect under time-reversal symmetry-breaking. The correlation between internal magnetism, Bogoliubov quasi-particles, and pairing modulation provides a chain of experimental clues for time-reversal symmetry-breaking kagome superconductivity., Comment: To appear in Nature Materials (2024)

Superconductivity involving finite momentum pairing can lead to spatial gap and pair density modulations, as well as Bogoliubov Fermi states within the superconducting gap. However, the experimental realization of their intertwined relations has been challenging. Here, we detect chiral kagome superconductivity modulations with residual Fermi arcs in KV3Sb5 and CsV3Sb5 by normal and Josephson scanning tunneling microscopy down to 30mK with resolved electronic energy difference at microelectronvolt level. We observe a U-shaped superconducting gap with flat residual in-gap states. This gap exhibits chiral 2 by 2 spatial modulations with magnetic field tunable chirality, which align with the chiral 2 by 2 pair density modulations observed through Josephson tunneling. These findings demonstrate a chiral pair density wave (PDW) that breaks time-reversal symmetry. Quasiparticle interference imaging of the in-gap zero-energy states reveals segmented arcs, with high-temperature data linking them to parts of the reconstructed V d-orbital states within the charge order. The detected residual Fermi arcs can be explained by the partial suppression of these d-orbital states through an interorbital 2 by 2 PDW and thus serve as candidate Bogoliubov Fermi states. Additionally, we differentiate the observed PDW order from impurity-induced gap modulations. Our observations not only uncover a chiral PDW order with orbital-selectivity, but also illuminate the fundamental space-momentum correspondence inherent in finite momentum paired superconductivity., Comment: To appear in Nature (2024)

The chiral 2 by 2 charge order has been reported and confirmed in the kagome superconductor RbV3Sb5, while its interplay with superconductivity remains elusive owing to its lowest superconducting transition temperature Tc of about 0.85K among the AV3Sb5 family (A=K, Rb, Cs) that severely challenges electronic spectroscopic probes. Here, utilizing dilution-refrigerator-based scanning tunneling microscopy (STM) down to 30mK, we observe chiral 2 by 2 pair density waves with residual Fermi arcs in RbV3Sb5. We find a superconducting gap of 150{\mu}eV with substantial residual in-gap states. The spatial distribution of this gap exhibits chiral 2 by 2 modulations, signaling a chiral pair density wave (PDW). Our quasi-particle interference imaging of the zero-energy residual states further reveals arc-like patterns. We discuss the relation of the gap modulations with the residual Fermi arcs under the space-momentum correspondence between PDW and Bogoliubov Fermi states., Comment: To appear in Chinese Physics Letters

Vertical federated learning (VFL) considers the case that the features of data samples are partitioned over different participants. VFL consists of two main steps, i.e., identify the common data samples for all participants (alignment) and train model using the aligned data samples (training). However, when there are many participants and data samples, both alignment and training become slow. As such, we propose TreeCSS as an efficient VFL framework that accelerates the two main steps. In particular, for sample alignment, we design an efficient multi-party private set intersection (MPSI) protocol called Tree-MPSI, which adopts a tree-based structure and a data-volume-aware scheduling strategy to parallelize alignment among the participants. As model training time scales with the number of data samples, we conduct coreset selection (CSS) to choose some representative data samples for training. Our CCS method adopts a clustering-based scheme for security and generality, which first clusters the features locally on each participant and then merges the local clustering results to select representative samples. In addition, we weight the samples according to their distances to the centroids to reflect their importance to model training. We evaluate the effectiveness and efficiency of our TreeCSS framework on various datasets and models. The results show that compared with vanilla VFL, TreeCSS accelerates training by up to 2.93x and achieves comparable model accuracy., Comment: 16 pages, 7 figures

In this paper, we present a decentralized network protocol, Space Network Protocol, based on Communication Satellite Services. The protocol outlines a method for distributing information about the status of satellite communication services across the entire blockchain network, facilitating fairness and transparency in all communication services. Our primary objective is to standardize the services delivered by all satellite networks under the communication satellite protocol. This standard remains intact regardless of potential unreliability associated with the satellites or the terminal hardware. We proposed PoD (Proof of Distribution) to verify if the communication satellites are online and PoF (Proof of Flow) to authenticate the actual data flow provided by the communication satellites. In addition, we also proposed PoM (Proof of Mesh) to verify if the communication satellites have successfully meshed together. Utilizing zero-knowledge proof and multi-party cryptographic computations, we can evaluate the service provisioning parameters of each satellite, even in the presence of potential terminal or network node fraud. This method offers technical support for the modeling of distributed network services.

Graph neural networks (GNNs) are machine learning models specialized for graph data and widely used in many applications. To train GNNs on large graphs that exceed CPU memory, several systems store data on disk and conduct out-of-core processing. However, these systems suffer from either read amplification when reading node features that are usually smaller than a disk page or degraded model accuracy by treating the graph as disconnected partitions. To close this gap, we build a system called DiskGNN, which achieves high I/O efficiency and thus fast training without hurting model accuracy. The key technique used by DiskGNN is offline sampling, which helps decouple graph sampling from model computation. In particular, by conducting graph sampling beforehand, DiskGNN acquires the node features that will be accessed by model computation, and such information is utilized to pack the target node features contiguously on disk to avoid read amplification. Besides, \name{} also adopts designs including four-level feature store to fully utilize the memory hierarchy to cache node features and reduce disk access, batched packing to accelerate the feature packing process, and pipelined training to overlap disk access with other operations. We compare DiskGNN with Ginex and MariusGNN, which are state-of-the-art systems for out-of-core GNN training. The results show that DiskGNN can speed up the baselines by over 8x while matching their best model accuracy.

Global popularity (GP) bias is the phenomenon that popular items are recommended much more frequently than they should be, which goes against the goal of providing personalized recommendations and harms user experience and recommendation accuracy. Many methods have been proposed to reduce GP bias but they fail to notice the fundamental problem of GP, i.e., it considers popularity from a \textit{global} perspective of \textit{all users} and uses a single set of popular items, and thus cannot capture the interests of individual users. As such, we propose a user-aware version of item popularity named \textit{personal popularity} (PP), which identifies different popular items for each user by considering the users that share similar interests. As PP models the preferences of individual users, it naturally helps to produce personalized recommendations and mitigate GP bias. To integrate PP into recommendation, we design a general \textit{personal popularity aware counterfactual} (PPAC) framework, which adapts easily to existing recommendation models. In particular, PPAC recognizes that PP and GP have both direct and indirect effects on recommendations and controls direct effects with counterfactual inference techniques for unbiased recommendations. All codes and datasets are available at \url{https://github.com/Stevenn9981/PPAC}., Comment: Accepted by WWW'24 as a research full paper

The Friedrich-Wintgen bound state in the continuum (FW BIC) provides a unique approach for achieving high quality factor (Q-factor) resonance, which has attracted wide attention and promoted the development of various applications. However, the FW BIC is usually considered as accident BIC resulting from the continuous parameters tuning, and a systematic approach to generate the FW BIC is still lacking. To address this, a method of actively forming FW BIC by matching the damping rate and resonance frequency of the coupling mode is proposed. As a proof-of-principle example, we propose a dielectric dimerized grating borophene heterostructure that generates a FW BIC near the commercially important communication wavelength. The coupling system comprises an electrically tunable borophene plasmon mode and a BIC supported by a dielectric dimer grating that can be attributed to the Brillouin zone folding. More interestingly, the BIC can be excited by the localized borophene plasmon (LBP) mode through near-field coupling as LBP mode can be considered as the dipole source. The interaction between them can further form the FW BIC, and support electromagnetically induced transparency (EIT)-like with maximum group index up to 2043, indicating its great potential for slow light applications. Our results provide a promising strategy and theoretical support for the generation of FW BIC in active plasmonic optical devices.

The decay of sunspot plays a key role in magnetic flux transportation in solar active regions (ARs). To better understand the physical mechanism of the entire decay process of a sunspot, an {\alpha}-configuration sunspot in AR NOAA 12411 was studied. Based on the continuum intensity images and vector magnetic field data with stray light correction from Solar Dynamics Observatory/Helioseismic and Magnetic Imager, the area, vector magnetic field and magnetic flux in the umbra and penumbra are calculated with time, respectively. Our main results are as follows: (1) The decay curves of the sunspot area in its umbra, penumbra, and whole sunspot take the appearance of Gaussian profiles. The area decay rates of the umbra, penumbra and whole sunspot are -1.56 MSH/day, -12.61 MSH/day and -14.04 MSH/day, respectively; (2) With the decay of the sunspot, the total magnetic field strength and the vertical component of the penumbra increase, and the magnetic field of the penumbra becomes more vertical. Meanwhile, the total magnetic field strength and vertical magnetic field strength for the umbra decrease, and the inclination angle changes slightly with an average value of about 20{\deg}; (3) The magnetic flux decay curves of the sunspot in its umbra, penumbra, and whole sunspot exhibit quadratic patterns, their magnetic flux decay rates of the umbra, penumbra and whole sunspot are -9.84 * 10^19 Mx/day, -1.59 * 10^20 Mx/day and -2.60 * 10^20 Mx/day , respectively. The observation suggests that the penumbra may be transformed into the umbra, resulting in the increase of the average vertical magnetic field strength and the reduction of the inclination angle in the penumbra during the decay of the sunspot.

Large Language Models (LLMs) are transforming the way people generate, explore, and engage with content. We study how we can develop LLM applications for online social networks. Despite LLMs' successes in other domains, it is challenging to develop LLM-based products for social networks for numerous reasons, and it has been relatively under-reported in the research community. We categorize LLM applications for social networks into three categories. First is knowledge tasks where users want to find new knowledge and information, such as search and question-answering. Second is entertainment tasks where users want to consume interesting content, such as getting entertaining notification content. Third is foundational tasks that need to be done to moderate and operate the social networks, such as content annotation and LLM monitoring. For each task, we share the challenges we found, solutions we developed, and lessons we learned. To the best of our knowledge, this is the first comprehensive paper about developing LLM applications for social networks.

Many applications involve the phenomenon of a material absorbing electromagnetic radiation. By exploiting wave interference, the efficiency of absorption can be significantly enhanced. Here, we propose Friedrich-Wintgen bound states in the continuum (F-W BICs) based on borophene metamaterials to realize coherent perfect absorption with a dual-band absorption peak in commercially important communication bands. The metamaterials consist of borophene gratings and a borophene sheet that can simultaneously support a Fabry-Perot plasmon resonance and a guided plasmon mode. The formation and dynamic modulation of the F-W BIC can be achieved by adjusting the width or carrier density of the borophene grating, while the strong coupling leads to the anti-crossover behavior of the absorption spectrum. Due to the weak angular dispersion originating from the intrinsic flat-band characteristic of the deep sub-wavelength periodic structure, the proposed plasmonic system exhibits almost no change in wavelength and absorption at large incident angles (within 70 degrees). In addition, we employ the temporal coupled-mode theory including near- and far-field coupling to obtain strong critical coupling, successfully achieve coherent perfect absorption, and can realize the absorption switch by changing the phase difference between the two coherent beams. Our findings can offer theoretical support for absorber design and all-optical tuning.

While generative AI excels in content generation, it does not always increase user engagement. This can be attributed to two main factors. First, generative AI generates content without incorporating explicit or implicit feedback about user interactions. Even if the generated content seems to be more informative or well-written, it does not necessarily lead to an increase in user activities, such as clicks. Second, there is a concern with the quality of the content generative AI produces, which often lacks the distinctiveness and authenticity that human-created content possesses. These two factors can lead to content that fails to meet specific needs and preferences of users, ultimately reducing its potential to be engaging. This paper presents a generic framework of how to improve user engagement with generative AI by leveraging user feedback. Our solutions employ rejection sampling, a technique used in reinforcement learning, to boost engagement metrics. We leveraged the framework in the context of email notification subject lines generation for an online social network, and achieved significant engagement metric lift including +1% Session and +0.4% Weekly Active Users. We believe our work offers a universal framework that enhances user engagement with generative AI, particularly when standard generative AI reaches its limits in terms of enhancing content to be more captivating. To the best of our knowledge, this represents an early milestone in the industry's successful use of generative AI to enhance user engagement.

Transformer-based large language models (e.g., BERT and GPT) achieve great success, and fine-tuning, which tunes a pre-trained model on a task-specific dataset, is the standard practice to utilize these models for downstream tasks. However, Transformer fine-tuning has long running time and high memory consumption due to the large size of the models. We propose the SPT system to fine-tune Transformer-based models efficiently by introducing sparsity. We observe that the memory consumption of Transformer mainly comes from storing attention weights for multi-head attention (MHA), and the majority of running time is spent on feed-forward network (FFN). Thus, we design the sparse MHA module, which computes and stores only large attention weights to reduce memory consumption, and the routed FFN module, which dynamically activates a subset of model parameters for each token to reduce computation cost. We implement SPT on PyTorch and customize CUDA kernels to run sparse MHA and routed FFN efficiently. Specifically, we use product quantization to identify the large attention weights and compute attention via sparse matrix multiplication for sparse MHA. For routed FFN, we batch the tokens according to their activated model parameters for efficient computation. We conduct extensive experiments to evaluate SPT on various model configurations. The results show that SPT consistently outperforms well-optimized baselines, reducing the peak memory consumption by up to 50% and accelerating fine-tuning by up to 2.2x., Comment: Firstly submitted to VLDB November 1, 2023, rejection received on December 15, 2023

For graph self-supervised learning (GSSL), masked autoencoder (MAE) follows the generative paradigm and learns to reconstruct masked graph edges or node features. Contrastive Learning (CL) maximizes the similarity between augmented views of the same graph and is widely used for GSSL. However, MAE and CL are considered separately in existing works for GSSL. We observe that the MAE and CL paradigms are complementary and propose the graph contrastive masked autoencoder (GCMAE) framework to unify them. Specifically, by focusing on local edges or node features, MAE cannot capture global information of the graph and is sensitive to particular edges and features. On the contrary, CL excels in extracting global information because it considers the relation between graphs. As such, we equip GCMAE with an MAE branch and a CL branch, and the two branches share a common encoder, which allows the MAE branch to exploit the global information extracted by the CL branch. To force GCMAE to capture global graph structures, we train it to reconstruct the entire adjacency matrix instead of only the masked edges as in existing works. Moreover, a discrimination loss is proposed for feature reconstruction, which improves the disparity between node embeddings rather than reducing the reconstruction error to tackle the feature smoothing problem of MAE. We evaluate GCMAE on four popular graph tasks (i.e., node classification, node clustering, link prediction, and graph classification) and compare with 14 state-of-the-art baselines. The results show that GCMAE consistently provides good accuracy across these tasks, and the maximum accuracy improvement is up to 3.2% compared with the best-performing baseline.

Among the many variants of graph neural network (GNN) architectures capable of modeling data with cross-instance relations, an important subclass involves layers designed such that the forward pass iteratively reduces a graph-regularized energy function of interest. In this way, node embeddings produced at the output layer dually serve as both predictive features for solving downstream tasks (e.g., node classification) and energy function minimizers that inherit desirable inductive biases and interpretability. However, scaling GNN architectures constructed in this way remains challenging, in part because the convergence of the forward pass may involve models with considerable depth. To tackle this limitation, we propose a sampling-based energy function and scalable GNN layers that iteratively reduce it, guided by convergence guarantees in certain settings. We also instantiate a full GNN architecture based on these designs, and the model achieves competitive accuracy and scalability when applied to the largest publicly-available node classification benchmark exceeding 1TB in size.

Trip flow between areas is a fundamental metric for human mobility research. Given its identification with travel demand and its relevance for transportation and urban planning, many models have been developed for its estimation. These models focus on flow intensity, disregarding the information provided by the local mobility orientation. A field-theoretic approach can overcome this issue and handling both intensity and direction at once. Here we propose a general vector-field representation starting from individuals' trajectories valid for any type of mobility. By introducing four models of spatial exploration, we show how individuals' elections determine the mesoscopic properties of the mobility field. Distance optimization in long displacements and random-like local exploration are necessary to reproduce empirical field features observed in Chinese logistic data and in New York City Foursquare check-ins. Our framework is an essential tool to capture hidden symmetries in mesoscopic urban mobility, it establishes a benchmark to test the validity of mobility models and opens the doors to the use of field theory in a wide spectrum of applications., Comment: 13 pages, 8 figures, Appendices

Multi-domain recommendation (MDR) aims to provide recommendations for different domains (e.g., types of products) with overlapping users/items and is common for platforms such as Amazon, Facebook, and LinkedIn that host multiple services. Existing MDR models face two challenges: First, it is difficult to disentangle knowledge that generalizes across domains (e.g., a user likes cheap items) and knowledge specific to a single domain (e.g., a user likes blue clothing but not blue cars). Second, they have limited ability to transfer knowledge across domains with small overlaps. We propose a new MDR method named EDDA with two key components, i.e., embedding disentangling recommender and domain alignment, to tackle the two challenges respectively. In particular, the embedding disentangling recommender separates both the model and embedding for the inter-domain part and the intra-domain part, while most existing MDR methods only focus on model-level disentangling. The domain alignment leverages random walks from graph processing to identify similar user/item pairs from different domains and encourages similar user/item pairs to have similar embeddings, enhancing knowledge transfer. We compare EDDA with 12 state-of-the-art baselines on 3 real datasets. The results show that EDDA consistently outperforms the baselines on all datasets and domains. All datasets and codes are available at https://github.com/Stevenn9981/EDDA., Comment: Accepted by CIKM'23 as a Long paper

Quantification of the overall characteristics of urban mobility using coarse-grained methods is crucial for urban management, planning and sustainable development. Although some recent studies have provided quantification methods for coarse-grained numerical information regarding urban mobility, a method that can simultaneously capture numerical and spatial information remains an outstanding problem. Here, we use mathematical vectors to depict human mobility, with mobility magnitude representing numerical information and mobility direction representing spatial information. We then define anisotropy and centripetality metrics by vector computation to measure imbalance in direction distribution and orientation toward the city center of mobility flows, respectively. As a case study, we apply our method to 60 Chinese cities and identify three mobility patterns: strong monocentric, weak monocentric and polycentric. To better understand mobility pattern, we further study the allometric scaling of the average commuting distance and the spatiotemporal variations of the two metrics in different patterns. Finally, we build a microscopic model to explain the key mechanisms driving the diversity in anisotropy and centripetality. Our work offers a comprehensive method that considers both numerical and spatial information to quantify and classify the overall characteristics of urban mobility, enhancing our understanding of the structure and evolution of urban mobility systems.