Recently, Hyper-relational Knowledge Graphs (HKGs) have been proposed as an extension of traditional Knowledge Graphs (KGs) to better represent real-world facts with additional qualifiers. As a result, researchers have attempted to adapt classical Knowledge Graph Embedding (KGE) models for HKGs by designing extra qualifier processing modules. However, it remains unclear whether the superior performance of Hyper-relational KGE (HKGE) models arises from their base KGE model or the specially designed extension module. In this paper, we data-wise convert HKGs to KG format using decomposition methods and then evaluate several classical KGE models' performance on HKGs. Our results show that some KGE models achieve comparable performance to HKGE models. Upon further analysis, we find that the decomposition methods alter the original HKG topology and fail to fully preserve HKG information. Moreover, we observe that current HKGE models are either insufficient in capturing the graph's long-range dependency or struggle to integrate main-triple and qualifier information due to the information compression issue. To further justify our findings and provide a direction for HKGE research, we propose FormerGNN, which employs a qualifier integrator to preserve the original HKG topology, a GNN-based graph encoder to capture the graph's long-range dependencies, and an improved approach for integrating main-triple and qualifier information to mitigate compression issues. Our experimental results demonstrate that FormerGNN outperforms existing HKGE models.

The semantic annotation of tabular data plays a crucial role in various downstream tasks. Previous research has proposed knowledge graph (KG)-based and deep learning-based methods, each with its inherent limitations. KG-based methods encounter difficulties annotating columns when there is no match for column cells in the KG. Moreover, KG-based methods can provide multiple predictions for one column, making it challenging to determine the semantic type with the most suitable granularity for the dataset. This type granularity issue limits their scalability. On the other hand, deep learning-based methods face challenges related to the valuable context missing issue. This occurs when the information within the table is insufficient for determining the correct column type. This paper presents KGLink, a method that combines Wiki-Data KG information with a pre-trained deep learning language model for table column annotation, effectively addressing both type granularity and valuable context missing issues. Through comprehensive experiments on widely used tabular datasets encompassing numeric and string columns with varying type granularity, we showcase the effectiveness and efficiency of KGLink. By leveraging the strengths of KGLink, we successfully surmount challenges related to type granularity and valuable context issues, establishing it as a robust solution for the semantic annotation of tabular data.

The rapidly changing nature of information world-wide often leads to incomplete and obsolete knowledge facts stored in knowledge bases (KBs). Therefore, reasoning over the dynamic KB sequences, which targets at knowledge inference from evolving facts, is of great importance to maintain KB completeness as well as freshness. Existing approaches for KB updating mainly either focus on knowledge representation learning methods, which suffer from lack of interpretability, or attempt to mine path-based logical rules, which are limited in capturing structural semantics of KB. In this work, we present KartGPS, a system for KB updating taking advantage of temporal graph pattern-based semantic (tGPS) rules. Specifically, the tGPS rules are learned from KB sequences and thus are capable of capturing both temporal and topological regularities of KBs along the evolving of time. Due to the huge amount and imperfect quality of tGPS rules, directly generating and applying all generated rules in a brute-force manner for knowledge updating over large-scale KB sequences would be highly time-consuming and error-prone. Therefore, we investigate the problem of Knowledge Update Rule Discovery (KURD), which aims at deriving an optimal subset of tGPS rules for performing knowledge updating, considering the rule quality and coverage. We show that the KURD problem is NP-hard and design two effective approximation algorithms with greedy and pruning strategies. We demonstrate the effectiveness and efficiency of proposed approaches by extensive experiments on real-world KB datasets.

Retrieval-Augmented Generation (RAG) has recently emerged as a promising solution to alleviate Large Language Model (LLM)’s deficiency in lack of knowledge. Existing RAG datasets, however, do not adequately represent the diverse and dynamic nature of real-world Question Answering (QA) tasks. To bridge this gap, we introduce the Comprehensive RAG Benchmark (CRAG), a factual question answering benchmark of 4,409 question-answer pairs and mock APIs to simulate web and Knowledge Graph (KG) search. CRAG is designed to encapsulate a diverse array of questions across five domains and eight question categories, reflecting varied entity popularity from popular to long-tail, and temporal dynamisms ranging from years to seconds. Our evaluation on this benchmark highlights the gap to fully trustworthy QA. Whereas most advanced LLMs achieve $\le 34\%$ accuracy on CRAG, adding RAG in a straightforward manner improves the accuracy only to 44%. State-of-the-art industry RAG solutions only answer 63% questions without any hallucination. CRAG also reveals much lower accuracy in answering questions regarding facts with higher dynamism, lower popularity, or higher complexity, suggesting future research directions. The CRAG benchmark laid the groundwork for a KDD Cup 2024 challenge, attracted thousands of participants and submissions. We commit to maintaining CRAG to serve research communities in advancing RAG solutions and general QA solutions. CRAG is available at https://github.com/facebookresearch/CRAG/.

In recent years, domain-specific knowledge bases (KBs) have attracted more attention in academics and industry because of their expertise and in-depth representation in a specific domain. However, when constructing a domain-specific KB, one needs to address not only the challenges in constructing a general KB, but also the difficulties raised by the nature of domain-specific raw data. Considering the usability of financial Knowledge Bases (KBs) in many downstream applications, such as financial risk analysis and fraud detection, we aim to build a Temporal Financial KB. However, the complex, time-varying financial knowledge and the volatile financial market evolution make construction quite challenging. Thus, we propose a Platform for Temporal Financial KB Construction, T-FinKB Platform1. This platform consists of three fundamental modules, i.e., evolved knowledge extraction module, temporal record linkage and conflicts resolution module, and dynamic knowledge update module designed for financial knowledge. Generated temporal financial KBs (T-FinKBs) from T-FinKB Platform can be updated automatically. T-FinKB has been successfully applied to a downstream application, stock trend prediction with backtesting evaluation. In addition, T-FinKB Platform provides several APIs for visualization, queries and analytics.

Understanding the semantics of tabular data is of great importance in various downstream applications, such as schema matching, data cleaning, and data integration. Column semantic type annotation is a critical task in the semantic understanding of tabular data. Despite the fact that various approaches have been proposed, they are challenged by the difficulties of handling wide tables and incorporating complex inter-table context information. Failure to handle wide tables limits the usage of column type annotation approaches, while failure to incorporate inter-table context harms the annotation quality. Existing methods either completely ignore these problems or propose ad-hoc solutions. In this paper, we propose Related tables Enhanced Column semantic type Annotation framework (RECA), which incorporates inter-table context information by finding and aligning schema-similar and topic-relevant tables based on a novel named entity schema. The design of RECA can naturally handle wide tables and incorporate useful inter-table context information to enhance the annotation quality. We conduct extensive experiments on two web table datasets to comprehensively evaluate the performance of RECA. Our results show that RECA achieves support-weighted F1 scores of 0.853 and 0.937 with macro average F1 scores of 0.674 and 0.783 on the two datasets respectively, which outperform the state-of-the-art methods.

Image-to-image regression is an important computer vision task. In this paper, we propose a novel image-to-image regression model following the recent trend in generative modeling that employs Stochastic Differential Equations (SDEs) and score matching. We first apply diffusion processes to regression data using designed SDEs, and then perform inference by gradually reversing the processes. In particular, our method uses synchronized diffusion, which simultaneously applies diffusion to both input and response images to stabilize diffusion and subsequent parameter learning. Furthermore, based on the Expectation-Maximization (EM) algorithm, we develop an effective algorithm for prediction. We implement a conditional U-Net architecture with pre-trained DenseNet encoder for our proposed model and refer to it as DenseSocre. Our new model is able to generate diverse outcomes for image colorization, and the proposed prediction algorithm is able to achieve close to state-of-art performance on high-resolution monocular depth estimation.

Heterogeneous information networks (HINs) are usually used to model information systems with multi-type objects and relations. In contrast, graphs that have a single type of nodes and edges, are often called homogeneous graphs. Measuring similarities among objects is an important task in data mining applications, such as web search, link prediction, and clustering. Currently, several similarity measures are defined for HINs. Most of these measures are based on meta-paths, which show sequences of node classes and edge types along the paths between two nodes. However, meta-paths, which are often designed by domain experts, are hard to enumerate and choose w.r.t. the quality of similarity scores. This makes using existing similarity measures in real applications difficult. To address this problem, we extend SimRank, a well-known similarity measure on homogeneous graphs, to HINs, by introducing the concept of the decay graph. The newly proposed similarity measure is called HowSim, which has the property of being meta-path free, and capturing the structural and semantic similarity simultaneously. The generality and effectiveness of HowSim, and the efficiency of our proposed algorithms for computing HowSim scores, are demonstrated by extensive experiments.

Knowledge extraction has become a hot topic recently with the increasing number of applications needed for large-scale knowledge bases (KBs), such as semantic search and QA systems. The goal of knowledge extraction is to extract relations and their arguments from natural language text. Recent research proposes two kinds of solutions. The first one, called Closed IE, tries to construct KB through predefined features or rules with respect to a specific domain. It requires specifying the interested predicates in advance, which restricts its application to the domains where prior knowledge about the interested predicates must be given. The second one, called Open IE, tries to extract facts by using the parsing structure from the unstructured text. However, they cannot avoid extracting redundant facts. Such extractions can hardly be directly used to populate the existing KB. Moreover, many correct extractions are not relevant to the document, which limits the applications to understand the essential information that the document conveys. In this paper, we propose an end-to-end system which takes a target incomplete KB and documents as input. It first performs joint entity and relation linking to the existing KB based on both contexts of document and background KB information. Then it summarizes the extracted facts by considering the relevance to the document and the diversity between them. Extensive experiments over real datasets demonstrate the effectiveness and efficiency of the proposed methods.

Image to image regression is an essential machine learning problem that is commonly encountered in computer vision tasks. Previous research works have been concentrating on task-dependent end-to-end regression models. Although such a discriminative approach has shown its proficiency in producing precise predictions, its architecture is usually overly complicated for interpretation and unable to provide uncertainty quantification for prediction. We propose a general image to image regression framework named CSMNet which is based on conditional score matching and Langevin dynamics. The proposed framework is generative and emphasize on the probabilistic perspective of general image to image regression. Discriminative models are also involved as sufficient feature extractors to complement the generative modeling of the distribution of the target image conditioned an input image. The proposed model is trained by multi-level denoising score matching. Subsequently, the learned model is used for sampling from the conditional distribution via annealed Langevin dynamics. We introduce a grid selection procedure of Langevin sampling which can produce robust prediction. Comprehensive experiments are conducted on diverse image to image regression problems. The proposed model surpasses other purely generative models and achieves comparable performance to some end-to-end regression methods. Our method also demonstrates remarkable generalization ability which further demonstrates the advantage of the proposed conditional score matching framework.

Nowadays, most openly available knowledge bases (KBs) are incomplete, since they are not synchronized with the emerging facts happening in the real world. Therefore, knowledge base population (KBP) from external data sources, which extracts knowledge from unstructured text to populate KBs, becomes a vital task. Recent research proposes two types of solutions that partially address this problem, but the performance of these solutions is limited. The first solution, dynamic KB construction from unstructured text, requires specifications of which predicates are of interest to the KB, which needs preliminary setups and is not suitable for an in-time population scenario. The second solution, Open Information Extraction (Open IE) from unstructured text, has limitations in producing facts that can be directly linked to the target KB without redundancy and ambiguity. In this paper, we present an end-to-end system, KBPearl, for KBP, which takes an incomplete KB and a large corpus of text as input, to (1) organize the noisy extraction from Open IE into canonicalized facts; and (2) populate the KB by joint entity and relation linking, utilizing the context knowledge of the facts and the side information inferred from the source text. We demonstrate the effectiveness and efficiency of KBPearl against the state-of-the-art techniques, through extensive experiments on real-world datasets.

With the scaling of technology, the computing industry is experiencing a shift from multi-core to many-core architectures. However, traditional metallic-based on-chip interconnects may not scale to support many-core architectures due to high power dissipation, and increased communication latency. Attention has recently shifted to emerging technologies such as silicon-photonics and wireless interconnects to implement future on-chip communications. Although emerging technologies show promising results for power-efficient, low-latency, and scalable on-chip interconnects, the use of single technology may not be sufficient to scale future architectures. In this paper, we extend the heterogeneous architecture Optical-Wireless Network-on-Chip (OWN [1]) to Reconfigurable Optical-Wireless Network-on-Chip (R-OWN) by introducing run-time reconfigurable wireless channels. Like OWN, R-OWN is designed such that one-hop photonic interconnect is used up to 64 cores (called a cluster) and communication beyond a cluster is one-hop wireless to limit the network diameter to a maximum of three hops. The photonic bandwidth is efficiently shared using time division multiplexing (TDM) while the wireless bandwidth is shared using frequency division multiplexing (FDM). By exploiting the heterogeneity of two emerging technologies, we reduce the energy/bit, improve performance via reconfiguration, and thereby improve the sustainability of NoCs and CMPs. We propose a preliminary assessment of implementing heterogeneous technologies with the router microarchitecture. Further, we also discuss the design of horn antenna for implementing the wireless channels. Our results indicate that R-OWN improves the performance (throughput and latency) by 15 percent when compared to OWN while consuming 7 percent more energy than OWN. Further, OWN and R-OWN improve energy-efficiency by 54-61 percent when compared to WCube and CMesh architectures, respectively. It should be noted that both OWN and R-OWN require less area than state-of-the-art wired, wireless, and optical on-chip networks.

Knowledge base construction (KBC) has become a hot and in-time topic recently with the increasing application need of large-scale knowledge bases (KBs), such as semantic search, QA systems, the Google Knowledge Graph and IBM Watson QA System. Existing KBs mainly focus on encoding the factual facts of the world, e.g., city area and company product, which are regarded as the objective knowledge, whereas the subjective knowledge, which is frequently mentioned in Web queries, has been neglected. The subjective knowledge has no documented ground truth, instead, the truth relies on people's dominant opinion, which can be solicited from online crowd workers. In our work, we propose a KBC framework for subjective knowledge base construction taking advantage of the knowledge from the crowd and existing KBs. We develop a two-staged framework for subjective KB construction which consists of core subjective KB construction and subjective KB enrichment. Firstly, we try to build a core subjective KB mined from existing KBs, where every instance has rich objective properties. Then, we populate the core subjective KB with instances extracted from existing KBs, in which the crowd is leverage to annotate the subjective property of the instances. In order to optimize the crowd annotation process, we formulate the problem of subjective KB enrichment procedure as a cost-aware instance annotation problem and propose two instance annotation algorithms, i.e., adaptive instance annotation and batch-mode instance annotation algorithms. We develop a two-stage system for subjective KB construction which consists of core subjective KB construction and subjective knowledge enrichment. We evaluate our framework on real knowledge bases and a real crowdsourcing platform, the experimental results show that we can derive high quality subjective knowledge facts from existing KBs and crowdsourcing techniques through our proposed framework.

Word embeddings have attracted much attention recently.Different from alphabetic writing systems, Chinese characters are often composed of subcharacter components which are also semantically informative.In this work, we propose an approach to jointly embed Chinese words as well as their characters and fine-grained subcharacter components.We use three likelihoods to evaluate whether the context words, characters, and components can predict the current target word, and collected 13,253 subcharacter components to demonstrate the existing approaches of decomposing Chinese characters are not enough.Evaluation on both word similarity and word analogy tasks demonstrates the superior performance of our model.

Ridesharing enables drivers to share any empty seats in their vehicles with riders to improve the efficiency of transportation for the benefit of both drivers and riders. Different from existing studies in ridesharing that focus on minimizing the travel costs of vehicles, we consider that the satisfaction of riders (the utility values) is more important nowadays. Thus, we formulate the problem of utility-aware ridesharing on road networks (URR) with the goal of providing the optimal rider schedules for vehicles to maximize the overall utility, subject to spatial-temporal and capacity constraints. To assign a new rider to a given vehicle, we propose an efficient algorithm with a minimum increase in travel cost without reordering the existing schedule of the vehicle. We prove that the URR problem is NP-hard by reducing it from the 0-1 Knapsack problem and it is unlikely to be approximated within any constant factor in polynomial time through a reduction from the DENS k-SUBGRAPH problem. Therefore, we propose three efficient approximate algorithms, including a bilateral arrangement algorithm, an efficient greedy algorithm and a grouping-based scheduling algorithm, to assign riders to suitable vehicles with a high overall utility. Through extensive experiments, we demonstrate the efficiency and effectiveness of our URR approaches on both real and synthetic data sets.

3-D additive manufacturing (AM) offers unprecedented flexibility in the realization of complicated 3-D structures. Polymer jetting is one of the promising 3-D AM techniques that utilizes photosensitive polymers as the build material and is capable of precisely printing electromagnetic (EM) components up into the THz range. In this paper, important design and implementation aspects of polymer-jetting-based 3-D-printed EM components are discussed. A number of 3-D-printable polymer materials and their broadband EM characterization from GHz to THz are introduced. Design methodologies specific for 3-D-printed antennas and other EM components are presented. As examples, various 3-D-printed devices operating from GHz to THz frequency, including electromagnetic crystals (EMXT), waveguide, horn antenna, gradient index (GRIN) lenses, as well as 3-D AM-enabled new designs, such as millimeter wave (mmW)/THz, reflect array antennas, computer-generated THz holograms, and so on are reviewed. Moreover, current limitations and possible future improvements of the polymer jetting technique for EM applications are discussed. This type of 3-D AM technique is likely to enable many novel antenna and circuit architectures as well as various interesting 3-D metamaterial structures.