arXiv subjects

Wei Hu

Publications and source records attributed to Wei Hu.

At least 199 records · Page 11Linked to original sources

arXiv2022

Representation Alignment in Neural Networks

It is now a standard for neural network representations to be trained on large, publicly available datasets, and used for new problems. The reasons for why neural network representations have been so successful for transfer, however, are still not fully understood. In this paper we show that, after training, neural network representations align their top singular vectors to the targets. We investigate this representation alignment phenomenon in a variety of neural network architectures and find that (a) alignment emerges across a variety of different architectures and optimizers, with more alignment arising from depth (b) alignment increases for layers closer to the output and (c) existing high-performance deep CNNs exhibit high levels of alignment. We then highlight why alignment between the top singular vectors and the targets can speed up learning and show in a classic synthetic transfer problem that representation alignment correlates with positive and negative transfer to similar and dissimilar tasks.

cs.LG↗

arXiv2022

I Know What You Do Not Know: Knowledge Graph Embedding via Co-distillation Learning

Knowledge graph (KG) embedding seeks to learn vector representations for entities and relations. Conventional models reason over graph structures, but they suffer from the issues of graph incompleteness and long-tail entities. Recent studies have used pre-trained language models to learn embeddings based on the textual information of entities and relations, but they cannot take advantage of graph structures. In the paper, we show empirically that these two kinds of features are complementary for KG embedding. To this end, we propose CoLE, a Co-distillation Learning method for KG Embedding that exploits the complementarity of graph structures and text information. Its graph embedding model employs Transformer to reconstruct the representation of an entity from its neighborhood subgraph. Its text embedding model uses a pre-trained language model to generate entity representations from the soft prompts of their names, descriptions, and relational neighbors. To let the two model promote each other, we propose co-distillation learning that allows them to distill selective knowledge from each other's prediction logits. In our co-distillation learning, each model serves as both a teacher and a student. Experiments on benchmark datasets demonstrate that the two models outperform their related baselines, and the ensemble method CoLE with co-distillation learning advances the state-of-the-art of KG embedding.

cs.CL↗

arXiv2022

Inductive Knowledge Graph Reasoning for Multi-batch Emerging Entities

Over the years, reasoning over knowledge graphs (KGs), which aims to infer new conclusions from known facts, has mostly focused on static KGs. The unceasing growth of knowledge in real life raises the necessity to enable the inductive reasoning ability on expanding KGs. Existing inductive work assumes that new entities all emerge once in a batch, which oversimplifies the real scenario that new entities continually appear. This study dives into a more realistic and challenging setting where new entities emerge in multiple batches. We propose a walk-based inductive reasoning model to tackle the new setting. Specifically, a graph convolutional network with adaptive relation aggregation is designed to encode and update entities using their neighboring relations. To capture the varying neighbor importance, we employ a query-aware feedback attention mechanism during the aggregation. Furthermore, to alleviate the sparse link problem of new entities, we propose a link augmentation strategy to add trustworthy facts into KGs. We construct three new datasets for simulating this multi-batch emergence scenario. The experimental results show that our proposed model outperforms state-of-the-art embedding-based, walk-based and rule-based models on inductive KG reasoning.

cs.CL↗

arXiv2022

Ferro-rotational domain walls revealed by electric quadrupole second harmonic generation microscopy

Domain walls are ubiquitous in materials that undergo phase transitions driven by spontaneous symmetry breaking. Domain walls in ferroics and multiferroics have received tremendous attention recently due to their emergent properties distinct from their domain counterparts, for example, their high mobility and controllability, as well as their potential applications in nanoelectronics. However, it is extremely challenging to detect, visualize and study the ferro-rotational (FR) domain walls because the FR order, in contrast to ferromagnetism (FM) and ferroelectricity (FE), is invariant under both the spatial-inversion and the time-reversal operations and thus hardly couple with conventional experimental probes. Here, an FR candidate $\mathrm{NiTiO_{3}}$ is investigated by ultrasensitive electric quadrupole (EQ) second harmonic generation rotational anisotropy (SHG RA) to probe the point symmetries of the two degenerate FR domain states, showing their relation by the vertical mirror operations that are broken below the FR critical temperature. We then visualize the real-space FR domains by scanning EQ SHG microscopy, and further resolve the FR domain walls by revealing a suppressed SHG intensity at domain walls. By taking local EQ SHG RA measurements, we show the restoration of the mirror symmetry at FR domain walls and prove their unconventional nonpolar nature. Our findings not only provide a comprehensive insight into FR domain walls, but also demonstrate a unique and powerful tool for future studies on domain walls of unconventional ferroics, both of which pave the way towards future manipulations and applications of FR domain walls.

cond-mat.mtrl-sci↗

arXiv2022

Large-scale Entity Alignment via Knowledge Graph Merging, Partitioning and Embedding

Entity alignment is a crucial task in knowledge graph fusion. However, most entity alignment approaches have the scalability problem. Recent methods address this issue by dividing large KGs into small blocks for embedding and alignment learning in each. However, such a partitioning and learning process results in an excessive loss of structure and alignment. Therefore, in this work, we propose a scalable GNN-based entity alignment approach to reduce the structure and alignment loss from three perspectives. First, we propose a centrality-based subgraph generation algorithm to recall some landmark entities serving as the bridges between different subgraphs. Second, we introduce self-supervised entity reconstruction to recover entity representations from incomplete neighborhood subgraphs, and design cross-subgraph negative sampling to incorporate entities from other subgraphs in alignment learning. Third, during the inference process, we merge the embeddings of subgraphs to make a single space for alignment search. Experimental results on the benchmark OpenEA dataset and the proposed large DBpedia1M dataset verify the effectiveness of our approach.

cs.LG↗

arXiv2022

Neural Capture of Animatable 3D Human from Monocular Video

We present a novel paradigm of building an animatable 3D human representation from a monocular video input, such that it can be rendered in any unseen poses and views. Our method is based on a dynamic Neural Radiance Field (NeRF) rigged by a mesh-based parametric 3D human model serving as a geometry proxy. Previous methods usually rely on multi-view videos or accurate 3D geometry information as additional inputs; besides, most methods suffer from degraded quality when generalized to unseen poses. We identify that the key to generalization is a good input embedding for querying dynamic NeRF: A good input embedding should define an injective mapping in the full volumetric space, guided by surface mesh deformation under pose variation. Based on this observation, we propose to embed the input query with its relationship to local surface regions spanned by a set of geodesic nearest neighbors on mesh vertices. By including both position and relative distance information, our embedding defines a distance-preserved deformation mapping and generalizes well to unseen poses. To reduce the dependency on additional inputs, we first initialize per-frame 3D meshes using off-the-shelf tools and then propose a pipeline to jointly optimize NeRF and refine the initial mesh. Extensive experiments show our method can synthesize plausible human rendering results under unseen poses and views.

cs.CV↗

arXiv2022

Rigidity degrees of indecomposable modules over representation-finite self-injective algebras

The rigidity degree of a generator-cogenerator determines the dominant dimension of its endomorphism algebra, and is closely related to a recently introduced homological dimension -- rigidity dimension. In this paper, we give explicit formulae for the rigidity degrees of all indecomposable modules over representation-finite self-injective algebras by developing combinatorial methods from the Euclidean algorithm. As an application, the rigidity dimensions of some algebras of types $A$ and $E$ are given.

math.RT↗

arXiv2022

KSSOLV 2.0: An efficient MATLAB toolbox for solving the Kohn-Sham equations with plane-wave basis set

KSSOLV (Kohn-Sham Solver) is a MATLAB toolbox for performing Kohn-Sham density functional theory (DFT) calculations with a plane-wave basis set. KSSOLV 2.0 preserves the design features of the original KSSOLV software to allow users and developers to easily set up a problem and perform ground-state calculations as well as to prototype and test new algorithms. Furthermore, it includes new functionalities such as new iterative diagonalization algorithms, k-point sampling for electron band structures, geometry optimization and advanced algorithms for performing DFT calculations with local, semi-local, and hybrid exchange-correlation functionals. It can be used to study the electronic structures of both molecules and solids. We describe these new capabilities in this work through a few use cases. We also demonstrate the numerical accuracy and computational efficiency of KSSOLV on a variety of examples.

physics.comp-ph↗

arXiv2022

$μ\text{KG}$: A Library for Multi-source Knowledge Graph Embeddings and Applications

This paper presents $μ\text{KG}$, an open-source Python library for representation learning over knowledge graphs. $μ\text{KG}$ supports joint representation learning over multi-source knowledge graphs (and also a single knowledge graph), multiple deep learning libraries (PyTorch and TensorFlow2), multiple embedding tasks (link prediction, entity alignment, entity typing, and multi-source link prediction), and multiple parallel computing modes (multi-process and multi-GPU computing). It currently implements 26 popular knowledge graph embedding models and supports 16 benchmark datasets. $μ\text{KG}$ provides advanced implementations of embedding techniques with simplified pipelines of different tasks. It also comes with high-quality documentation for ease of use. $μ\text{KG}$ is more comprehensive than existing knowledge graph embedding libraries. It is useful for a thorough comparison and analysis of various embedding models and tasks. We show that the jointly learned embeddings can greatly help knowledge-powered downstream tasks, such as multi-hop knowledge graph question answering. We will stay abreast of the latest developments in the related fields and incorporate them into $μ\text{KG}$.

cs.CL↗

arXiv2022

Skimming, Locating, then Perusing: A Human-Like Framework for Natural Language Video Localization

This paper addresses the problem of natural language video localization (NLVL). Almost all existing works follow the "only look once" framework that exploits a single model to directly capture the complex cross- and self-modal relations among video-query pairs and retrieve the relevant segment. However, we argue that these methods have overlooked two indispensable characteristics of an ideal localization method: 1) Frame-differentiable: considering the imbalance of positive/negative video frames, it is effective to highlight positive frames and weaken negative ones during the localization. 2) Boundary-precise: to predict the exact segment boundary, the model should capture more fine-grained differences between consecutive frames since their variations are often smooth. To this end, inspired by how humans perceive and localize a segment, we propose a two-step human-like framework called Skimming-Locating-Perusing (SLP). SLP consists of a Skimming-and-Locating (SL) module and a Bi-directional Perusing (BP) module. The SL module first refers to the query semantic and selects the best matched frame from the video while filtering out irrelevant frames. Then, the BP module constructs an initial segment based on this frame, and dynamically updates it by exploring its adjacent frames until no frame shares the same activity semantic. Experimental results on three challenging benchmarks show that our SLP is superior to the state-of-the-art methods and localizes more precise segment boundaries.

cs.CV↗

arXiv2022

Reducing the Vision and Language Bias for Temporal Sentence Grounding

Temporal sentence grounding (TSG) is an important yet challenging task in multimedia information retrieval. Although previous TSG methods have achieved decent performance, they tend to capture the selection biases of frequently appeared video-query pairs in the dataset rather than present robust multimodal reasoning abilities, especially for the rarely appeared pairs. In this paper, we study the above issue of selection biases and accordingly propose a Debiasing-TSG (D-TSG) model to filter and remove the negative biases in both vision and language modalities for enhancing the model generalization ability. Specifically, we propose to alleviate the issue from two perspectives: 1) Feature distillation. We built a multi-modal debiasing branch to firstly capture the vision and language biases, and then apply a bias identification module to explicitly recognize the true negative biases and remove them from the benign multi-modal representations. 2) Contrastive sample generation. We construct two types of negative samples to enforce the model to accurately learn the aligned multi-modal semantics and make complete semantic reasoning. We apply the proposed model to both commonly and rarely appeared TSG cases, and demonstrate its effectiveness by achieving the state-of-the-art performance on three benchmark datasets (ActivityNet Caption, TACoS, and Charades-STA).

cs.CV↗

arXiv2022

What you get is not always what you see: pitfalls in solar array assessment using overhead imagery

Effective integration planning for small, distributed solar photovoltaic (PV) arrays into electric power grids requires access to high quality data: the location and power capacity of individual solar PV arrays. Unfortunately, national databases of small-scale solar PV do not exist; those that do are limited in their spatial resolution, typically aggregated up to state or national levels. While several promising approaches for solar PV detection have been published, strategies for evaluating the performance of these models are often highly heterogeneous from study to study. The resulting comparison of these methods for practical applications for energy assessments becomes challenging and may imply that the reported performance evaluations are overly optimistic. The heterogeneity comes in many forms, each of which we explore in this work: the level of spatial aggregation, the validation of ground truth, inconsistencies in the training and validation datasets, and the degree of diversity of the locations and sensors from which the training and validation data originate. For each, we discuss emerging practices from the literature to address them or suggest directions of future research. As part of our investigation, we evaluate solar PV identification performance in two large regions. Our findings suggest that traditional performance evaluation of the automated identification of solar PV from satellite imagery may be optimistic due to common limitations in the validation process. The takeaways from this work are intended to inform and catalyze the large-scale practical application of automated solar PV assessment techniques by energy researchers and professionals.

cs.CV↗

arXiv2022

Enhancing Document-level Relation Extraction by Entity Knowledge Injection

Document-level relation extraction (RE) aims to identify the relations between entities throughout an entire document. It needs complex reasoning skills to synthesize various knowledge such as coreferences and commonsense. Large-scale knowledge graphs (KGs) contain a wealth of real-world facts, and can provide valuable knowledge to document-level RE. In this paper, we propose an entity knowledge injection framework to enhance current document-level RE models. Specifically, we introduce coreference distillation to inject coreference knowledge, endowing an RE model with the more general capability of coreference reasoning. We also employ representation reconciliation to inject factual knowledge and aggregate KG representations and document representations into a unified space. The experiments on two benchmark datasets validate the generalization of our entity knowledge injection framework and the consistent improvement to several document-level RE models.

cs.CL↗

arXiv2022

Facing Changes: Continual Entity Alignment for Growing Knowledge Graphs

Entity alignment is a basic and vital technique in knowledge graph (KG) integration. Over the years, research on entity alignment has resided on the assumption that KGs are static, which neglects the nature of growth of real-world KGs. As KGs grow, previous alignment results face the need to be revisited while new entity alignment waits to be discovered. In this paper, we propose and dive into a realistic yet unexplored setting, referred to as continual entity alignment. To avoid retraining an entire model on the whole KGs whenever new entities and triples come, we present a continual alignment method for this task. It reconstructs an entity's representation based on entity adjacency, enabling it to generate embeddings for new entities quickly and inductively using their existing neighbors. It selects and replays partial pre-aligned entity pairs to train only parts of KGs while extracting trustworthy alignment for knowledge augmentation. As growing KGs inevitably contain non-matchable entities, different from previous works, the proposed method employs bidirectional nearest neighbor matching to find new entity alignment and update old alignment. Furthermore, we also construct new datasets by simulating the growth of multilingual DBpedia. Extensive experiments demonstrate that our continual alignment method is more effective than baselines based on retraining or inductive learning.

cs.CL↗

arXiv2022

Brief Industry Paper: The Necessity of Adaptive Data Fusion in Infrastructure-Augmented Autonomous Driving System

This paper is the first to provide a thorough system design overview along with the fusion methods selection criteria of a real-world cooperative autonomous driving system, named Infrastructure-Augmented Autonomous Driving or IAAD. We present an in-depth introduction of the IAAD hardware and software on both road-side and vehicle-side computing and communication platforms. We extensively characterize the IAAD system in the context of real-world deployment scenarios and observe that the network condition that fluctuates along the road is currently the main technical roadblock for cooperative autonomous driving. To address this challenge, we propose new fusion methods, dubbed "inter-frame fusion" and "planning fusion" to complement the current state-of-the-art "intra-frame fusion". We demonstrate that each fusion method has its own benefit and constraint.

cs.DC↗

arXiv2022

Unidirectional vortex waveguides and multistable vortex pairs in polariton condensates

Vortices carrying quantized topological charges have potential application in information processing. In this work, we investigate vortex carriers and waveguides in microcavity polariton condensates, nonresonantly excited by a homogeneous pump with intensity grooves. An intensity groove with a ring shape in the pump gives rise to dark-ring states of the condensate with a $π$-phase jump, akin to dark solitons. The dark-ring states can be destroyed by a stronger density of the surrounding condensate and reduce into vortex-antivortex pairs. Multiple vortex-pair states are found to be stable in the same dark ring of the pump. When the pump ring is broader, higher-order dark states with multiple $π$-phase jumps can be obtained and interestingly they can be used to construct vortex waveguides. If a single vortex is imprinted in such waveguides, it can travel in a particular direction, showing one-way transportation. In other words, an imprinted vortex with a certain charge in a specifically designed higher-order dark state is only allowed to propagate unidirectionally.

physics.optics↗

arXiv2022

Thermal radiative cooling of carbon cluster cations C$_N^+$, $N = 9, 11,12, 17-27$

The radiative cooling rates of C$_N^+$ clusters ($N = 9, 11, 12, 17-27$) have been measured in the ultrahigh vacuum of an electrostatic storage ring to values on the order of $10^4$ s$^{-1}$. The rates were measured as a competing channel to unimolecular decay, and the rate constants pertain to the excitation energies where these two channels compete. Such high values can only be explained as photon emission from thermally excited electronic states, a mechanism that has also been seen in polycyclic aromatic hydrocarbon cations. The high rates have a very strong stabilizing effect on the clusters and the underlying mechanism gives a high energy conversion efficiency, with the potential to reach high quantum efficiencies in the emission process. The competing decay of unimolecular fragmentation defines upper limits for photon energies that can be down-converted to lower energy photons. Including previously measured cluster sizes provides the limits for all clusters C$_N^+$, $N=8-27$, of values that vary from 10 to 14.5 eV, with a general increase with size. Clusters absorbing photons of energies below these limits cool down efficiently by emission of photons via electronic transitions and their fragmentation is strongly reduced, increasing their survival in HI regions.

physics.chem-ph↗

arXiv2022

Dynamic Point Cloud Denoising via Gradient Fields

3D dynamic point clouds provide a discrete representation of real-world objects or scenes in motion, which have been widely applied in immersive telepresence, autonomous driving, surveillance, etc. However, point clouds acquired from sensors are usually perturbed by noise, which affects downstream tasks such as surface reconstruction and analysis. Although many efforts have been made for static point cloud denoising, dynamic point cloud denoising remains under-explored. In this paper, we propose a novel gradient-field-based dynamic point cloud denoising method, exploiting the temporal correspondence via the estimation of gradient fields -- a fundamental problem in dynamic point cloud processing and analysis. The gradient field is the gradient of the log-probability function of the noisy point cloud, based on which we perform gradient ascent so as to converge each point to the underlying clean surface. We estimate the gradient of each surface patch and exploit the temporal correspondence, where the temporally corresponding patches are searched leveraging on rigid motion in classical mechanics. In particular, we treat each patch as a rigid object, which moves in the gradient field of an adjacent frame via force until reaching a balanced state, i.e., when the sum of gradients over the patch reaches 0. Since the gradient would be smaller when the point is closer to the underlying surface, the balanced patch would fit the underlying surface well, thus leading to the temporal correspondence. Finally, the position of each point in the patch is updated along the direction of the gradient averaged from corresponding patches in adjacent frames. Experimental results demonstrate that the proposed model outperforms state-of-the-art methods under both synthetic noise and simulated real-world noise.

cs.CV↗