arXiv subjects

Haiyang Yu

Publications and source records attributed to Haiyang Yu.

At least 127 records · Page 7Linked to original sources

arXiv2023

Universal Information Extraction with Meta-Pretrained Self-Retrieval

Universal Information Extraction~(Universal IE) aims to solve different extraction tasks in a uniform text-to-structure generation manner. Such a generation procedure tends to struggle when there exist complex information structures to be extracted. Retrieving knowledge from external knowledge bases may help models to overcome this problem but it is impossible to construct a knowledge base suitable for various IE tasks. Inspired by the fact that large amount of knowledge are stored in the pretrained language models~(PLM) and can be retrieved explicitly, in this paper, we propose MetaRetriever to retrieve task-specific knowledge from PLMs to enhance universal IE. As different IE tasks need different knowledge, we further propose a Meta-Pretraining Algorithm which allows MetaRetriever to quicktly achieve maximum task-specific retrieval performance when fine-tuning on downstream IE tasks. Experimental results show that MetaRetriever achieves the new state-of-the-art on 4 IE tasks, 12 datasets under fully-supervised, low-resource and few-shot scenarios.

cs.CL↗

arXiv2023

FAST reveals new evidence for M94 as a merger

We report the first high-sensitivity HI observation toward the spiral galaxy M94 with the Five-hundred-meter Aperture Spherical radio Telescope (FAST). From these observations, we discovered that M94 has a very extended HI disk, twice larger than that observed by THINGS, which is accompanied by an HI filament and seven HVCs (high velocity clouds) at different distances. The projected distances of these clouds and filament are less than 50 kpc from the galactic center. We measured a total integrated flux (including all clouds/filament) of 127.3 ($\pm$1) Jy km s$^{-1}$, corresponding to a H I mass of (6.51$\pm$0.06)$\times$10$^{8}$M$_{\odot}$, which is 63.0% more than that observed by THINGS. By comparing numerical simulations with the HI maps and the optical morphology of M94, we suggest that M94 is likely a remnant of a major merger of two galaxies, and the HVCs and HI filament could be the tidal features originated from the first collision of the merger happened about 5 Gyr ago. Furthermore, we found a seemingly isolated HI cloud at a projection distance of 109 kpc without any optical counterpart detected. We discussed the possibilities of the origin of this cloud, such as dark dwarf galaxy and RELHIC (REionization-Limited HI Cloud). Our results demonstrate that high-sensitivity and wide-field HI imaging is important in revealing the diffuse cold gas structures and tidal debris which is crucial to understanding the dynamical evolution of galaxies.

astro-ph.GA↗

arXiv2023

Energy landscape reveals the underlying mechanism of cancer-adipose conversion with gene network models

Cancer is a systemic heterogeneous disease involving complex molecular networks. Tumor formation involves epithelial-mesenchymal transition (EMT), which promotes both metastasis and plasticity of cancer cells. Recent experiments proposed that cancer cells can be transformed into adipocytes with combination drugs. However, the underlying mechanisms for how these drugs work from molecular network perspective remain elusive. To reveal the mechanism of cancer-adipose conversion (CAC), we adopt a systems biology approach by combing mathematical modeling and molecular experiments based on the underlying molecular regulatory network. We identified four types of attractors which correspond to epithelial (E), mesenchymal (M), adipose (A) and partial/intermediate EMT (P) cell states on the CAC landscape. Landscape and transition path results illustrate that the intermediate states play critical roles in cancer to adipose transition. Through a landscape control strategy, we identified two new therapeutic strategies for drug combinations to promote CAC. We further verified these predictions by molecular experiments in different cell lines. Our combined computational and experimental approach provides a powerful tool to explore molecular mechanisms for cell fate transitions in cancer networks. Our results revealed the underlying mechanism for intermediate cell states governing the CAC, and identified new potential drug combinations to induce cancer adipogenesis.

q-bio.MN↗

arXiv2023

FAST discovery of long tidal tails in NGC 4490/85

We report the discovery of a 100 kpc HI tail in the merging galaxy pair NGC 4490/85 detected by the Five-Hundred-meter Aperture Spherical radio Telescope (FAST). The tidal tails extended in both the south and north directions, and they are much longer than that reported previously based on the VLA interferometric maps. The NGC 4490/85 is surrounded by a large gas envelope, and a starburst low metallicity dwarf galaxy MAPS 1231+42 is found to be connected with the gas envelope, indicating that galaxy interaction trigged the intense star formation in it. Based on the fact that the metallicity in MAPS 1231+42 is one order of magnitude lower than that in the two disks of NGC 4490 and NGC 4485, we speculate that the gas near this galaxy should be primordial and could be due to gas inflow from the circum-galactic medium (CGM). We also found a collimated gas component pointing at a nearby dwarf galaxy KK 149, suggesting that this galaxy might also be interacting with the NGC 4490 pair. We discuss the possible origin of the long tidal tails and the extended gas envelope in this merging system based on the new data from FAST.

astro-ph.GA↗

arXiv2023

Weakly-Supervised Text Instance Segmentation

Text segmentation is a challenging vision task with many downstream applications. Current text segmentation methods require pixel-level annotations, which are expensive in the cost of human labor and limited in application scenarios. In this paper, we take the first attempt to perform weakly-supervised text instance segmentation by bridging text recognition and text segmentation. The insight is that text recognition methods provide precise attention position of each text instance, and the attention location can feed to both a text adaptive refinement head (TAR) and a text segmentation head. Specifically, the proposed TAR generates pseudo labels by performing two-stage iterative refinement operations on the attention location to fit the accurate boundaries of the corresponding text instance. Meanwhile, the text segmentation head takes the rough attention location to predict segmentation masks which are supervised by the aforementioned pseudo labels. In addition, we design a mask-augmented contrastive learning by treating our segmentation result as an augmented version of the input text image, thus improving the visual representation and further enhancing the performance of both recognition and segmentation. The experimental results demonstrate that the proposed method significantly outperforms weakly-supervised instance segmentation methods on ICDAR13-FST (18.95$\%$ improvement) and TextSeg (17.80$\%$ improvement) benchmarks.

cs.CV↗

arXiv2023

Uncertainty-aware U-Net for Medical Landmark Detection

Heatmap-based methods play an important role in anatomical landmark detection. However, most current heatmap-based methods assume that the distributions of all landmarks are the same and the distribution of each landmark is isotropic, which may not be in line with reality. For example, the landmark on the jaw is more likely to be located along the edge and less likely to be located inside or outside the jaw. Manually annotating tends to follow similar rules, resulting in an anisotropic distribution for annotated landmarks, which represents the uncertainty in the annotation. To estimate the uncertainty, we propose a module named Pyramid Covariance Predictor to predict the covariance matrices of the target Gaussian distributions, which determine the distributions of landmarks and represent the uncertainty of landmark annotation. Specifically, the Pyramid Covariance Predictor utilizes the pyramid features extracted by the encoder of the backbone U-Net and predicts the Cholesky decomposition of the covariance matrix of the landmark location distribution. Experimental results show that the proposed Pyramid Covariance Predictor can accurately predict the distributions and improve the performance of anatomical landmark detection.

cs.CV↗

arXiv2023

Coarse-to-Fine Knowledge Selection for Document Grounded Dialogs

Multi-document grounded dialogue systems (DGDS) belong to a class of conversational agents that answer users' requests by finding supporting knowledge from a collection of documents. Most previous studies aim to improve the knowledge retrieval model or propose more effective ways to incorporate external knowledge into a parametric generation model. These methods, however, focus on retrieving knowledge from mono-granularity language units (e.g. passages, sentences, or spans in documents), which is not enough to effectively and efficiently capture precise knowledge in long documents. This paper proposes Re3G, which aims to optimize both coarse-grained knowledge retrieval and fine-grained knowledge extraction in a unified framework. Specifically, the former efficiently finds relevant passages in a retrieval-and-reranking process, whereas the latter effectively extracts finer-grain spans within those passages to incorporate into a parametric answer generation model (BART, T5). Experiments on DialDoc Shared Task demonstrate the effectiveness of our method.

cs.CL↗

arXiv2023

High sensitivity HI image of diffuse gas and new tidal features in M51 observed by FAST

We observed the classical interacting galaxy M51 with FAST and obtain high sensitivity HI image with column density down to 3.8 $\times$ 10$^{18}$ cm$^{-2}$. In the image we can see a diffuse extended envelope around the system and several new tidal features. We also get a deeper look at M51b's probable gas, which has an approximated velocity range of 560 to 740 km s$^{-1}$ and a flux of 7.5 Jy km s$^{-1}$. Compared to the VLA image, we observe more complete structures of the Southeast Tail, Northeast Cloud and Northwest Plume, as well as new features of the Northwest Cloud and Southwest Plume. M51's most prominent tidal feature, the Southeast Tail, looks very long and broad, in addition with two small detached clouds at the periphery. Due to the presence of optical and simulated counterparts, the Northwest cloud appears to be the tail of M51a, while the Northwest Plume is more likely a tidal tail of M51b. The large mass of the Northwest Plume suggests that M51b may have been as gas-rich as M51a before the interaction. In addition, the formation process of the Northeast Cloud and Southwest Plume is obscured by the lack of optical and simulated counterparts. These novel tidal features, together with M51b's probable gas, will inspire future simulations and provide a deeper understanding of the evolution of this interacting system.

astro-ph.GA↗

arXiv2022

Towards Generalized Open Information Extraction

Open Information Extraction (OpenIE) facilitates the open-domain discovery of textual facts. However, the prevailing solutions evaluate OpenIE models on in-domain test sets aside from the training corpus, which certainly violates the initial task principle of domain-independence. In this paper, we propose to advance OpenIE towards a more realistic scenario: generalizing over unseen target domains with different data distributions from the source training domains, termed Generalized OpenIE. For this purpose, we first introduce GLOBE, a large-scale human-annotated multi-domain OpenIE benchmark, to examine the robustness of recent OpenIE models to domain shifts, and the relative performance degradation of up to 70% implies the challenges of generalized OpenIE. Then, we propose DragonIE, which explores a minimalist graph expression of textual fact: directed acyclic graph, to improve the OpenIE generalization. Extensive experiments demonstrate that DragonIE beats the previous methods in both in-domain and out-of-domain settings by as much as 6.0% in F1 score absolutely, but there is still ample room for improvement.

cs.CL↗

arXiv2022

Benchmarking Chinese Text Recognition: Datasets, Baselines, and an Empirical Study

The flourishing blossom of deep learning has witnessed the rapid development of text recognition in recent years. However, the existing text recognition methods are mainly proposed for English texts. As another widely-spoken language, Chinese text recognition (CTR) in all ways has extensive application markets. Based on our observations, we attribute the scarce attention on CTR to the lack of reasonable dataset construction standards, unified evaluation protocols, and results of the existing baselines. To fill this gap, we manually collect CTR datasets from publicly available competitions, projects, and papers. According to application scenarios, we divide the collected datasets into four categories including scene, web, document, and handwriting datasets. Besides, we standardize the evaluation protocols in CTR. With unified evaluation protocols, we evaluate a series of representative text recognition methods on the collected datasets to provide baselines. The experimental results indicate that the performance of baselines on CTR datasets is not as good as that on English datasets due to the characteristics of Chinese texts that are quite different from the Latin alphabet. Moreover, we observe that by introducing radical-level supervision as an auxiliary task, the performance of baselines can be further boosted. The code and datasets are made publicly available at https://github.com/FudanVI/benchmarking-chinese-text-recognition

cs.CV↗

arXiv2022

Prompt Conditioned VAE: Enhancing Generative Replay for Lifelong Learning in Task-Oriented Dialogue

Lifelong learning (LL) is vital for advanced task-oriented dialogue (ToD) systems. To address the catastrophic forgetting issue of LL, generative replay methods are widely employed to consolidate past knowledge with generated pseudo samples. However, most existing generative replay methods use only a single task-specific token to control their models. This scheme is usually not strong enough to constrain the generative model due to insufficient information involved. In this paper, we propose a novel method, prompt conditioned VAE for lifelong learning (PCLL), to enhance generative replay by incorporating tasks' statistics. PCLL captures task-specific distributions with a conditional variational autoencoder, conditioned on natural language prompts to guide the pseudo-sample generation. Moreover, it leverages a distillation process to further consolidate past knowledge by alleviating the noise in pseudo samples. Experiments on natural language understanding tasks of ToD systems demonstrate that PCLL significantly outperforms competitive baselines in building LL models.

cs.CL↗

arXiv2022

Chinese Character Recognition with Radical-Structured Stroke Trees

The flourishing blossom of deep learning has witnessed the rapid development of Chinese character recognition. However, it remains a great challenge that the characters for testing may have different distributions from those of the training dataset. Existing methods based on a single-level representation (character-level, radical-level, or stroke-level) may be either too sensitive to distribution changes (e.g., induced by blurring, occlusion, and zero-shot problems) or too tolerant to one-to-many ambiguities. In this paper, we represent each Chinese character as a stroke tree, which is organized according to its radical structures, to fully exploit the merits of both radical and stroke levels in a decent way. We propose a two-stage decomposition framework, where a Feature-to-Radical Decoder perceives radical structures and radical regions, and a Radical-to-Stroke Decoder further predicts the stroke sequences according to the features of radical regions. The generated radical structures and stroke sequences are encoded as a Radical-Structured Stroke Tree (RSST), which is fed to a Tree-to-Character Translator based on the proposed Weighted Edit Distance to match the closest candidate character in the RSST lexicon. Our extensive experimental results demonstrate that the proposed method outperforms the state-of-the-art single-level methods by increasing margins as the distribution difference becomes more severe in the blurring, occlusion, and zero-shot scenarios, which indeed validates the robustness of the proposed method.

cs.CV↗

arXiv2022

Semi-Supervised Lifelong Language Learning

Lifelong learning aims to accumulate knowledge and alleviate catastrophic forgetting when learning tasks sequentially. However, existing lifelong language learning methods only focus on the supervised learning setting. Unlabeled data, which can be easily accessed in real-world scenarios, are underexplored. In this paper, we explore a novel setting, semi-supervised lifelong language learning (SSLL), where a model learns sequentially arriving language tasks with both labeled and unlabeled data. We propose an unlabeled data enhanced lifelong learner to explore SSLL. Specially, we dedicate task-specific modules to alleviate catastrophic forgetting and design two modules to exploit unlabeled data: (1) a virtual supervision enhanced task solver is constructed on a teacher-student framework to mine the underlying knowledge from unlabeled data; and (2) a backward augmented learner is built to encourage knowledge transfer from newly arrived unlabeled data to previous tasks. Experimental results on various language tasks demonstrate our model's effectiveness and superiority over competitive baselines under the new setting SSLL.

cs.CL↗

arXiv2022

Doc2Bot: Accessing Heterogeneous Documents via Conversational Bots

This paper introduces Doc2Bot, a novel dataset for building machines that help users seek information via conversations. This is of particular interest for companies and organizations that own a large number of manuals or instruction books. Despite its potential, the nature of our task poses several challenges: (1) documents contain various structures that hinder the ability of machines to comprehend, and (2) user information needs are often underspecified. Compared to prior datasets that either focus on a single structural type or overlook the role of questioning to uncover user needs, the Doc2Bot dataset is developed to target such challenges systematically. Our dataset contains over 100,000 turns based on Chinese documents from five domains, larger than any prior document-grounded dialog dataset for information seeking. We propose three tasks in Doc2Bot: (1) dialog state tracking to track user intentions, (2) dialog policy learning to plan system actions and contents, and (3) response generation which generates responses based on the outputs of the dialog policy. Baseline methods based on the latest deep learning models are presented, indicating that our proposed tasks are challenging and worthy of further research.

cs.CL↗

arXiv2022

Layout-Aware Information Extraction for Document-Grounded Dialogue: Dataset, Method and Demonstration

Building document-grounded dialogue systems have received growing interest as documents convey a wealth of human knowledge and commonly exist in enterprises. Wherein, how to comprehend and retrieve information from documents is a challenging research problem. Previous work ignores the visual property of documents and treats them as plain text, resulting in incomplete modality. In this paper, we propose a Layout-aware document-level Information Extraction dataset, LIE, to facilitate the study of extracting both structural and semantic knowledge from visually rich documents (VRDs), so as to generate accurate responses in dialogue systems. LIE contains 62k annotations of three extraction tasks from 4,061 pages in product and official documents, becoming the largest VRD-based information extraction dataset to the best of our knowledge. We also develop benchmark methods that extend the token-based language model to consider layout features like humans. Empirical results show that layout is critical for VRD-based extraction, and system demonstration also verifies that the extracted knowledge can help locate the answers that users care about.

cs.CL↗

arXiv2022

Explainability in Graph Neural Networks: A Taxonomic Survey

Deep learning methods are achieving ever-increasing performance on many artificial intelligence tasks. A major limitation of deep models is that they are not amenable to interpretability. This limitation can be circumvented by developing post hoc techniques to explain the predictions, giving rise to the area of explainability. Recently, explainability of deep models on images and texts has achieved significant progress. In the area of graph data, graph neural networks (GNNs) and their explainability are experiencing rapid developments. However, there is neither a unified treatment of GNN explainability methods, nor a standard benchmark and testbed for evaluations. In this survey, we provide a unified and taxonomic view of current GNN explainability methods. Our unified and taxonomic treatments of this subject shed lights on the commonalities and differences of existing methods and set the stage for further methodological developments. To facilitate evaluations, we generate a set of benchmark graph datasets specifically for GNN explainability. We summarize current datasets and metrics for evaluating GNN explainability. Altogether, this work provides a unified methodological treatment of GNN explainability and a standardized testbed for evaluations.

cs.LG↗

arXiv2022

A Survey on Neural Open Information Extraction: Current Status and Future Directions

Open Information Extraction (OpenIE) facilitates domain-independent discovery of relational facts from large corpora. The technique well suits many open-world natural language understanding scenarios, such as automatic knowledge base construction, open-domain question answering, and explicit reasoning. Thanks to the rapid development in deep learning technologies, numerous neural OpenIE architectures have been proposed and achieve considerable performance improvement. In this survey, we provide an extensive overview of the-state-of-the-art neural OpenIE models, their key design decisions, strengths and weakness. Then, we discuss limitations of current solutions and the open issues in OpenIE problem itself. Finally we list recent trends that could help expand its scope and applicability, setting up promising directions for future research in OpenIE. To our best knowledge, this paper is the first review on this specific topic.

cs.CL↗

arXiv2022

GraphFM: Improving Large-Scale GNN Training via Feature Momentum

Training of graph neural networks (GNNs) for large-scale node classification is challenging. A key difficulty lies in obtaining accurate hidden node representations while avoiding the neighborhood explosion problem. Here, we propose a new technique, named feature momentum (FM), that uses a momentum step to incorporate historical embeddings when updating feature representations. We develop two specific algorithms, known as GraphFM-IB and GraphFM-OB, that consider in-batch and out-of-batch data, respectively. GraphFM-IB applies FM to in-batch sampled data, while GraphFM-OB applies FM to out-of-batch data that are 1-hop neighborhood of in-batch data. We provide a convergence analysis for GraphFM-IB and some theoretical insight for GraphFM-OB. Empirically, we observe that GraphFM-IB can effectively alleviate the neighborhood explosion problem of existing methods. In addition, GraphFM-OB achieves promising performance on multiple large-scale graph datasets.

cs.LG↗