Search

Long-term motion generation is a challenging task that requires producing coherent and realistic sequences over extended durations. Current methods primarily rely on framewise motion representations, which capture only static spatial details and overlook temporal dynamics. This approach leads to significant redundancy across the temporal dimension, complicating the generation of effective long-term motion. To overcome these limitations, we introduce the novel concept of Lagrangian Motion Fields, specifically designed for long-term motion generation. By treating each joint as a Lagrangian particle with uniform velocity over short intervals, our approach condenses motion representations into a series of "supermotions" (analogous to superpixels). This method seamlessly integrates static spatial information with interpretable temporal dynamics, transcending the limitations of existing network architectures and motion sequence content types. Our solution is versatile and lightweight, eliminating the need for neural network preprocessing. Our approach excels in tasks such as long-term music-to-dance generation and text-to-motion generation, offering enhanced efficiency, superior generation quality, and greater diversity compared to existing methods. Additionally, the adaptability of Lagrangian Motion Fields extends to applications like infinite motion looping and fine-grained controlled motion generation, highlighting its broad utility. Video demonstrations are available at \url{https://plyfager.github.io/LaMoG}., Comment: 13 pages, 9 figures

Safety alignment is indispensable for Large language models (LLMs) to defend threats from malicious instructions. However, recent researches reveal safety-aligned LLMs prone to reject benign queries due to the exaggerated safety issue, limiting their helpfulness. In this paper, we propose a Safety-Conscious Activation Steering (SCANS) method to mitigate the exaggerated safety concerns in aligned LLMs. First, SCANS extracts the refusal steering vectors within the activation space and utilizes vocabulary projection to anchor some specific safety-critical layers which influence model refusal behavior. Second, by tracking the hidden state transition, SCANS identifies the steering direction and steers the model behavior accordingly, achieving a balance between exaggerated safety and adequate safety. Experiments show that SCANS achieves new state-of-the-art performance on XSTest and OKTest benchmarks, without impairing their defense capability against harmful queries and maintaining almost unchanged model capability.

Retrosynthesis analysis is pivotal yet challenging in drug discovery and organic chemistry. Despite the proliferation of computational tools over the past decade, AI-based systems often fall short in generalizing across diverse reaction types and exploring alternative synthetic pathways. This paper presents BatGPT-Chem, a large language model with 15 billion parameters, tailored for enhanced retrosynthesis prediction. Integrating chemical tasks via a unified framework of natural language and SMILES notation, this approach synthesizes extensive instructional data from an expansive chemical database. Employing both autoregressive and bidirectional training techniques across over one hundred million instances, BatGPT-Chem captures a broad spectrum of chemical knowledge, enabling precise prediction of reaction conditions and exhibiting strong zero-shot capabilities. Superior to existing AI methods, our model demonstrates significant advancements in generating effective strategies for complex molecules, as validated by stringent benchmark tests. BatGPT-Chem not only boosts the efficiency and creativity of retrosynthetic analysis but also establishes a new standard for computational tools in synthetic design. This development empowers chemists to adeptly address the synthesis of novel compounds, potentially expediting the innovation cycle in drug manufacturing and materials science. We release our trial platform at \url{https://www.batgpt.net/dapp/chem}.

Hyperbolic representations have shown remarkable efficacy in modeling inherent hierarchies and complexities within data structures. Hyperbolic neural networks have been commonly applied for learning such representations from data, but they often fall short in preserving the geometric structures of the original feature spaces. In response to this challenge, our work applies the Gromov-Wasserstein (GW) distance as a novel regularization mechanism within hyperbolic neural networks. The GW distance quantifies how well the original data structure is maintained after embedding the data in a hyperbolic space. Specifically, we explicitly treat the layers of the hyperbolic neural networks as a transport map and calculate the GW distance accordingly. We validate that the GW distance computed based on a training set well approximates the GW distance of the underlying data distribution. Our approach demonstrates consistent enhancements over current state-of-the-art methods across various tasks, including few-shot image classification, as well as semi-supervised graph link prediction and node classification., Comment: Accepted for ECCV 2024

Sparse Vector Coding (SVC) has long been considered an encoding method that meets the URLLC QOS requirements. This encoding method has been widely studied and applied due to its low encoding and decoding complexity, no pilot transmission, resistance to inter-carrier interference, and low power consumption. However, due to the use of position indexing, the encoding essentially reduces the signal-to-noise ratio requirements by increasing the communication bandwidth, which also leads to low encoding efficiency and strong rigidity in decoding. Based on the sparse representation characteristics of SVC, we propose a joint sparse representation encoding, namely Sparse Matrix Coding (SMC). This encoding method utilizes multi-user information joint encoding, and the sparsity and sparse locations between users are shared.

The joint optimization of the sensor trajectory and 3D map is a crucial characteristic of bundle adjustment (BA), essential for autonomous driving. This paper presents $\nu$-DBA, a novel framework implementing geometric dense bundle adjustment (DBA) using 3D neural implicit surfaces for map parametrization, which optimizes both the map surface and trajectory poses using geometric error guided by dense optical flow prediction. Additionally, we fine-tune the optical flow model with per-scene self-supervision to further improve the quality of the dense mapping. Our experimental results on multiple driving scene datasets demonstrate that our method achieves superior trajectory optimization and dense reconstruction accuracy. We also investigate the influences of photometric error and different neural geometric priors on the performance of surface reconstruction and novel view synthesis. Our method stands as a significant step towards leveraging neural implicit representations in dense bundle adjustment for more accurate trajectories and detailed environmental mapping.

Point-based interactive editing serves as an essential tool to complement the controllability of existing generative models. A concurrent work, DragDiffusion, updates the diffusion latent map in response to user inputs, causing global latent map alterations. This results in imprecise preservation of the original content and unsuccessful editing due to gradient vanishing. In contrast, we present DragNoise, offering robust and accelerated editing without retracing the latent map. The core rationale of DragNoise lies in utilizing the predicted noise output of each U-Net as a semantic editor. This approach is grounded in two critical observations: firstly, the bottleneck features of U-Net inherently possess semantically rich features ideal for interactive editing; secondly, high-level semantics, established early in the denoising process, show minimal variation in subsequent stages. Leveraging these insights, DragNoise edits diffusion semantics in a single denoising step and efficiently propagates these changes, ensuring stability and efficiency in diffusion editing. Comparative experiments reveal that DragNoise achieves superior control and semantic retention, reducing the optimization time by over 50% compared to DragDiffusion. Our codes are available at https://github.com/haofengl/DragNoise., Comment: Accepted by CVPR 2024

Webpage entity extraction is a fundamental natural language processing task in both research and applications. Nowadays, the majority of webpage entity extraction models are trained on structured datasets which strive to retain textual content and its structure information. However, existing datasets all overlook the rich hypertext features (e.g., font color, font size) which show their effectiveness in previous works. To this end, we first collect a \textbf{H}ypertext \textbf{E}ntity \textbf{E}xtraction \textbf{D}ataset (\textit{HEED}) from the e-commerce domains, scraping both the text and the corresponding explicit hypertext features with high-quality manual entity annotations. Furthermore, we present the \textbf{Mo}E-based \textbf{E}ntity \textbf{E}xtraction \textbf{F}ramework (\textit{MoEEF}), which efficiently integrates multiple features to enhance model performance by Mixture of Experts and outperforms strong baselines, including the state-of-the-art small-scale models and GPT-3.5-turbo. Moreover, the effectiveness of hypertext features in \textit{HEED} and several model components in \textit{MoEEF} are analyzed.

Large Language Models (LLMs) suffer from huge number of parameters, which restricts their deployment on edge devices. Weight sharing is one promising solution that encourages weight reuse, effectively reducing memory usage with less performance drop. However, current weight sharing techniques primarily focus on small-scale models like BERT and employ coarse-grained sharing rules, e.g., layer-wise. This becomes limiting given the prevalence of LLMs and sharing an entire layer or block obviously diminishes the flexibility of weight sharing. In this paper, we present a perspective on $\textit{$\textbf{head-wise shareable attention for large language models}$}$. We further propose two memory-efficient methods that share parameters across attention heads, with a specific focus on LLMs. Both of them use the same dynamic strategy to select the shared weight matrices. The first method directly reuses the pre-trained weights without retraining, denoted as $\textbf{DirectShare}$. The second method first post-trains with constraint on weight matrix similarity and then shares, denoted as $\textbf{PostShare}$. Experimental results reveal our head-wise shared models still maintain satisfactory capabilities, demonstrating the feasibility of fine-grained weight sharing applied to LLMs.

Large language models (LLMs) based on transformer are witnessing a notable trend of size expansion, which brings considerable costs to both model training and inference. However, existing methods such as model quantization, knowledge distillation, and model pruning are constrained by various issues, including hardware support limitations, the need for extensive training, and alterations to the internal structure of the model. In this paper, we propose a concise layer-wise pruning method called \textit{Layer Collapse (LaCo)}, in which rear model layers collapse into a prior layer, enabling a rapid reduction in model size while preserving the model structure. Comprehensive experiments show that our method maintains an average task performance of over 80\% at pruning ratios of 25-30\%, significantly outperforming existing state-of-the-art structured pruning methods. We also conduct post-training experiments to confirm that the proposed pruning method effectively inherits the parameters of the original model. Finally, we discuss our motivation from the perspective of layer-wise similarity and evaluate the performance of the pruned LLMs across various pruning ratios.

Time series modeling is uniquely challenged by the presence of autocorrelation in both historical and label sequences. Current research predominantly focuses on handling autocorrelation within the historical sequence but often neglects its presence in the label sequence. Specifically, emerging forecast models mainly conform to the direct forecast (DF) paradigm, generating multi-step forecasts under the assumption of conditional independence within the label sequence. This assumption disregards the inherent autocorrelation in the label sequence, thereby limiting the performance of DF-based models. In response to this gap, we introduce the Frequency-enhanced Direct Forecast (FreDF), which bypasses the complexity of label autocorrelation by learning to forecast in the frequency domain. Our experiments demonstrate that FreDF substantially outperforms existing state-of-the-art methods including iTransformer and is compatible with a variety of forecast models.

Network is a major bottleneck in modern cloud databases that adopt a storage-disaggregation architecture. Computation pushdown is a promising solution to tackle this issue, which offloads some computation tasks to the storage layer to reduce network traffic. Existing cloud OLAP systems statically decide whether to push down computation during the query optimization phase and do not consider the storage layer's computational capacity and load. Besides, there is a lack of a general principle that determines which operators are amenable for pushdown. Existing systems design and implement pushdown features empirically, which ends up picking a limited set of pushdown operators respectively. In this paper, we first design Adaptive pushdown as a new mechanism to avoid throttling the storage-layer computation during pushdown, which pushes the request back to the computation layer at runtime if the storage-layer computational resource is insufficient. Moreover, we derive a general principle to identify pushdown-amenable computational tasks, by summarizing common patterns of pushdown capabilities in existing systems. We propose two new pushdown operators, namely, selection bitmap and distributed data shuffle. Evaluation results on TPC-H show that Adaptive pushdown can achieve up to 1.9x speedup over both No pushdown and Eager pushdown baselines, and the new pushdown operators can further accelerate query execution by up to 3.0x., Comment: 13 pages, 15 figures

Detecting unusual patterns in graph data is a crucial task in data mining. However, existing methods face challenges in consistently achieving satisfactory performance and often lack interpretability, which hinders our understanding of anomaly detection decisions. In this paper, we propose a novel approach to graph anomaly detection that leverages the power of interpretability to enhance performance. Specifically, our method extracts an attention map derived from gradients of graph neural networks, which serves as a basis for scoring anomalies. Notably, our approach is flexible and can be used in various anomaly detection settings. In addition, we conduct theoretical analysis using synthetic data to validate our method and gain insights into its decision-making process. To demonstrate the effectiveness of our method, we extensively evaluate our approach against state-of-the-art graph anomaly detection techniques on real-world graph classification and wireless network datasets. The results consistently demonstrate the superior performance of our method compared to the baselines.

Generative adversarial networks (GANs) are popular for generative tasks; however, they often require careful architecture selection, extensive empirical tuning, and are prone to mode collapse. To overcome these challenges, we propose a novel model that identifies the low-dimensional structure of the underlying data distribution, maps it into a low-dimensional latent space while preserving the underlying geometry, and then optimally transports a reference measure to the embedded distribution. We prove three key properties of our method: 1) The encoder preserves the geometry of the underlying data; 2) The generator is $c$-cyclically monotone, where $c$ is an intrinsic embedding cost employed by the encoder; and 3) The discriminator's modulus of continuity improves with the geometric preservation of the data. Numerical experiments demonstrate the effectiveness of our approach in generating high-quality images and exhibiting robustness to both mode collapse and training instability., Comment: 21 pages excluding references

Recovering 3D geometry and textures of individual objects is crucial for many robotics applications, such as manipulation, pose estimation, and autonomous driving. However, decomposing a target object from a complex background is challenging. Most existing approaches rely on costly manual labels to acquire object instance perception. Recent advancements in 2D self-supervised learning offer new prospects for identifying objects of interest, yet leveraging such noisy 2D features for clean decomposition remains difficult. In this paper, we propose a Decomposed Object Reconstruction (DORec) network based on neural implicit representations. Our key idea is to use 2D self-supervised features to create two levels of masks for supervision: a binary mask for foreground regions and a K-cluster mask for semantically similar regions. These complementary masks result in robust decomposition. Experimental results on different datasets show DORec's superiority in segmenting and reconstructing diverse foreground objects from varied backgrounds enabling downstream tasks such as pose estimation.

While Large language models (LLMs) have garnered widespread applications across various domains due to their powerful language understanding and generation capabilities, the detection of non-factual or hallucinatory content generated by LLMs remains scarce. Currently, one significant challenge in hallucination detection is the laborious task of time-consuming and expensive manual annotation of the hallucinatory generation. To address this issue, this paper first introduces a method for automatically constructing model-specific hallucination datasets based on existing fact-checking datasets called AutoHall. Furthermore, we propose a zero-resource and black-box hallucination detection method based on self-contradiction. We conduct experiments towards prevalent open-/closed-source LLMs, achieving superior hallucination detection performance compared to extant baselines. Moreover, our experiments reveal variations in hallucination proportions and types among different models.

This paper presents predicate transfer, a novel method that optimizes join performance by pre-filtering tables to reduce the join input sizes. Predicate transfer generalizes Bloom join, which conducts pre-filtering within a single join operation, to multi-table joins such that the filtering benefits can be significantly increased. Predicate transfer is inspired by the seminal theoretical results by Yannakakis, which uses semi-joins to pre-filter acyclic queries. Predicate transfer generalizes the theoretical results to any join graphs and use Bloom filters to replace semi-joins leading to significant speedup. Evaluation shows predicate transfer can outperform Bloom join by 3.1x on average on TPC-H benchmark., Comment: 6 pages, 4 figures

BatGPT is a large-scale language model designed and trained jointly by Wuhan University and Shanghai Jiao Tong University. It is capable of generating highly natural and fluent text in response to various types of input, including text prompts, images, and audio. In the modeling level, we employ a bidirectional autoregressive architecture that allows the model to efficiently capture the complex dependencies of natural language, making it highly effective in tasks such as language generation, dialog systems, and question answering. Moreover, the bidirectional autoregressive modeling not only operates from left to right but also from right to left, effectively reducing fixed memory effects and alleviating model hallucinations. In the training aspect, we propose a novel parameter expansion method for leveraging the pre-training of smaller models and employ reinforcement learning from both AI and human feedback, aimed at improving the model's alignment performance. Overall, these approaches significantly improve the effectiveness of BatGPT, and the model can be utilized for a wide range of natural language applications.

As the capabilities of large language models (LLMs) continue to advance, evaluating their performance becomes increasingly crucial and challenging. This paper aims to bridge this gap by introducing CMMLU, a comprehensive Chinese benchmark that covers various subjects, including natural science, social sciences, engineering, and humanities. We conduct a thorough evaluation of 18 advanced multilingual- and Chinese-oriented LLMs, assessing their performance across different subjects and settings. The results reveal that most existing LLMs struggle to achieve an average accuracy of 50%, even when provided with in-context examples and chain-of-thought prompts, whereas the random baseline stands at 25%. This highlights significant room for improvement in LLMs. Additionally, we conduct extensive experiments to identify factors impacting the models' performance and propose directions for enhancing LLMs. CMMLU fills the gap in evaluating the knowledge and reasoning capabilities of large language models within the Chinese context.

Large Language Models (LLMs) have attained the impressive capability to resolve a wide range of NLP tasks by fine-tuning high-quality instruction data. However, collecting human-written data of high quality, especially multi-turn dialogues, is expensive and unattainable for most people. Though previous studies have used powerful LLMs to generate the dialogues automatically, they all suffer from generating untruthful dialogues because of the model hallucination. Therefore, we propose a method called RefGPT to generate enormous truthful and customized dialogues without worrying about factual errors caused by the model hallucination. RefGPT solves the model hallucination in dialogue generation by restricting the LLMs to leverage the given reference instead of reciting their own knowledge to generate dialogues. Additionally, RefGPT adds detailed controls on every utterance to enable high customization capability, which previous studies have ignored. On the basis of RefGPT, we also propose two high-quality dialogue datasets generated by GPT-4, namely RefGPT-Fact and RefGPT-Code. RefGPT-Fact is a dataset with 100k multi-turn dialogues based on factual knowledge and RefGPT-Code has 76k multi-turn dialogues covering a wide range of coding scenarios. Our code and datasets are released in https://github.com/mutonix/RefGPT., Comment: Accepted by EMNLP 2023

Topological semimetal materials have become a research hotspot due to their intrinsic strong spin-orbit coupling which leads to large charge-to-spin conversion efficiency and novel transport behaviors. In this work, we have observed a bilinear magnetoelectric resistance (BMER) of up to 0.1 nm2A-1Oe-1 in a singlelayer of sputtered semimetal Pt3Sn at room temperature. Different from previous observations, the value of BMER in sputtered Pt3Sn does not change out-of-plane due to the polycrystalline nature of Pt3Sn. The observation of BMER provides strong evidence of the existence of spin-momentum locking in the sputtered polycrystalline Pt3Sn. By adding an adjacent CoFeB magnetic layer, the BMER value of this bilayer system is doubled compared to the single Pt3Sn layer. This work broadens the material system in BMER study, which paves the way for the characterization of topological states and applications for spin memory and logic devices.

Named Entity Recognition (NER) is a cornerstone NLP task while its robustness has been given little attention. This paper rethinks the principles of NER attacks derived from sentence classification, as they can easily violate the label consistency between the original and adversarial NER examples. This is due to the fine-grained nature of NER, as even minor word changes in the sentence can result in the emergence or mutation of any entities, resulting in invalid adversarial examples. To this end, we propose a novel one-word modification NER attack based on a key insight, NER models are always vulnerable to the boundary position of an entity to make their decision. We thus strategically insert a new boundary into the sentence and trigger the Entity Boundary Interference that the victim model makes the wrong prediction either on this boundary word or on other words in the sentence. We call this attack Virtual Boundary Attack (ViBA), which is shown to be remarkably effective when attacking both English and Chinese models with a 70%-90% attack success rate on state-of-the-art language models (e.g. RoBERTa, DeBERTa) and also significantly faster than previous methods.

Generating photorealistic images with controllable camera pose and scene contents is essential for many applications including AR/VR and simulation. Despite the fact that rapid progress has been made in 3D-aware generative models, most existing methods focus on object-centric images and are not applicable to generating urban scenes for free camera viewpoint control and scene editing. To address this challenging task, we propose UrbanGIRAFFE, which uses a coarse 3D panoptic prior, including the layout distribution of uncountable stuff and countable objects, to guide a 3D-aware generative model. Our model is compositional and controllable as it breaks down the scene into stuff, objects, and sky. Using stuff prior in the form of semantic voxel grids, we build a conditioned stuff generator that effectively incorporates the coarse semantic and geometry information. The object layout prior further allows us to learn an object generator from cluttered scenes. With proper loss functions, our approach facilitates photorealistic 3D-aware image synthesis with diverse controllability, including large camera movement, stuff editing, and object manipulation. We validate the effectiveness of our model on both synthetic and real-world datasets, including the challenging KITTI-360 dataset., Comment: Project page: https://lv3d.github.io/urbanGIRAFFE

The classic three-terminal electronic transistors and the emerging two-terminal ion-based memristors are complementary to each other in various nonconventional information processing systems in a heterogeneous integration approach, such as hybrid CMOS/memristive neuromorphic crossbar arrays. Recent attempts to introduce transitive functions into memristors have given rise to gate-tunable memristive functions, hetero-plasticity and mixed-plasticity functions. However, it remains elusive under what application scenarios and in what ways transistors can benefit from the incorporation of ion-based memristive effects. Here, we introduce a new type of transistor named 'mempolar transistor' to the transistor family. Its polarity can be converted reversibly, in a nonvolatile fashion, between n-type and p-type depending on the history of the applied electrical stimulus. This is achieved by the use of the emerging semiconducting tellurium as the electrochemically active source/drain contact material, in combination with monolayer two-dimensional MoS2 channel, which results in a gated lateral Te/MoS2/Te memristor, or from a different perspective, a transistor whose channel can be converted reversibly between n-type MoS2 and p-type Te. With this unique mempolar function, our transistor holds the promise for reconfigurable logic circuits and secure circuits. In addition, we propose and demonstrate experimentally, a ternary content-addressable memory made of only two mempolar transistors, which used to require a dozen normal transistors, and by simulations, a device-inspired and hardware matched regularization method 'FlipWeight' for training artificial neural networks, which can achieve comparable performance to that achieved by the prevalent 'Dropout' and 'DropConnect' methods. This work represents a major advance in diversifying the functionality of transistors.

The understanding of neural networks as neuron-synapse binaries has been the foundation of neuroscience, and therefore, the emerging neuromorphic computing technology that takes inspiration from the brain. This dogma, however, has been increasingly challenged by recent neuroscience research in which the downplayed dendrites were found to be active, dynamically unique and computationally powerful. To date, research on artificial dendrites is scarce and the few existing devices are still far from versatile or (and) bio-realistic. A breakthrough is hampered by the limited available physical mechanisms in mainstream device architectures. Here, we experimentally demonstrate a bio-realistic and versatile artificial dendrite made of WSe2/MoS2 heterojunction-channel anti-ambipolar transistor, which can closely mimic the experimentally recorded non-monotonic dendritic Ca2+ action potential that underpins various sophisticated computations. Spiking neural network simulations reveal that the incorporation of this nonconventional but bio-realistic dendritic activation enhances the robustness and representational capability of the network in non-stationary environments. By further exploiting the memristive effect, dendritic anti-ambipolar transistor can naturally mimic Ca2+-mediated regulation of synaptic plasticity (meta-plasticity) at dendritic spine, an important homeostatic phenomenon due to dendritic modulation. The invention of dendritic anti-ambipolar transistor completes the family of neuromorphic transistors and represents a major advance in diversifying the functionality of transistors.

Detecting both known and unknown objects is a fundamental skill for robot manipulation in unstructured environments. Open-set object detection (OSOD) is a promising direction to handle the problem consisting of two subtasks: objects and background separation, and open-set object classification. In this paper, we present Openset RCNN to address the challenging OSOD. To disambiguate unknown objects and background in the first subtask, we propose to use classification-free region proposal network (CF-RPN) which estimates the objectness score of each region purely using cues from object's location and shape preventing overfitting to the training categories. To identify unknown objects in the second subtask, we propose to represent them using the complementary region of known categories in a latent space which is accomplished by a prototype learning network (PLN). PLN performs instance-level contrastive learning to encode proposals to a latent space and builds a compact region centering with a prototype for each known category. Further, we note that the detection performance of unknown objects can not be unbiasedly evaluated on the situation that commonly used object detection datasets are not fully annotated. Thus, a new benchmark is introduced by reorganizing GraspNet-1billion, a robotic grasp pose detection dataset with complete annotation. Extensive experiments demonstrate the merits of our method. We finally show that our Openset RCNN can endow the robot with an open-set perception ability to support robotic rearrangement tasks in cluttered environments. More details can be found in https://sites.google.com/view/openset-rcnn/, Comment: IEEE Robotics and Automation Letters

The recent rapid growth in mobile data traffic entails a pressing demand for improving the throughput of the underlying wireless communication networks. Network node deployment has been considered as an effective approach for throughput enhancement which, however, often leads to highly non-trivial non-convex optimizations. Although convex approximation based solutions are considered in the literature, their approximation to the actual throughput may be loose and sometimes lead to unsatisfactory performance. With this consideration, in this paper, we propose a novel graph neural network (GNN) method for the network node deployment problem. Specifically, we fit a GNN to the network throughput and use the gradients of this GNN to iteratively update the locations of the network nodes. Besides, we show that an expressive GNN has the capacity to approximate both the function value and the gradients of a multivariate permutation-invariant function, as a theoretic support to the proposed method. To further improve the throughput, we also study a hybrid node deployment method based on this approach. To train the desired GNN, we adopt a policy gradient algorithm to create datasets containing good training samples. Numerical experiments show that the proposed methods produce competitive results compared to the baselines.

As a fundamental natural language processing task and one of core knowledge extraction techniques, named entity recognition (NER) is widely used to extract information from texts for downstream tasks. Nested NER is a branch of NER in which the named entities (NEs) are nested with each other. However, most of the previous studies on nested NER usually apply linear structure to model the nested NEs which are actually accommodated in a hierarchical structure. Thus in order to address this mismatch, this work models the full nested NEs in a sentence as a holistic structure, then we propose a holistic structure parsing algorithm to disclose the entire NEs once for all. Besides, there is no research on applying corpus-level information to NER currently. To make up for the loss of this information, we introduce Point-wise Mutual Information (PMI) and other frequency features from corpus-aware statistics for even better performance by holistic modeling from sentence-level to corpus-level. Experiments show that our model yields promising results on widely-used benchmarks which approach or even achieve state-of-the-art. Further empirical studies show that our proposed corpus-aware features can substantially improve NER domain adaptation, which demonstrates the surprising advantage of our proposed corpus-level holistic structure modeling.