Search

The black-box nature of deep learning models in NLP hinders their widespread application. The research focus has shifted to Hierarchical Attribution (HA) for its ability to model feature interactions. Recent works model non-contiguous combinations with a time-costly greedy search in Eculidean spaces, neglecting underlying linguistic information in feature representations. In this work, we introduce a novel method, namely Poincar\'e Explanation (PE), for modeling feature interactions using hyperbolic spaces in an $O(n^2logn)$ time complexity. Inspired by Poincar\'e model, we propose a framework to project the embeddings into hyperbolic spaces, which exhibit better inductive biases for syntax and semantic hierarchical structures. Eventually, we prove that the hierarchical clustering process in the projected space could be viewed as building a minimum spanning tree and propose a time efficient algorithm. Experimental results demonstrate the effectiveness of our approach., Comment: 12 pages, 10 figures

KEPLMs are pre-trained models that utilize external knowledge to enhance language understanding. Previous language models facilitated knowledge acquisition by incorporating knowledge-related pre-training tasks learned from relation triples in knowledge graphs. However, these models do not prioritize learning embeddings for entity-related tokens. Moreover, updating the entire set of parameters in KEPLMs is computationally demanding. This paper introduces TRELM, a Robust and Efficient Pre-training framework for Knowledge-Enhanced Language Models. We observe that entities in text corpora usually follow the long-tail distribution, where the representations of some entities are suboptimally optimized and hinder the pre-training process for KEPLMs. To tackle this, we employ a robust approach to inject knowledge triples and employ a knowledge-augmented memory bank to capture valuable information. Furthermore, updating a small subset of neurons in the feed-forward networks (FFNs) that store factual knowledge is both sufficient and efficient. Specifically, we utilize dynamic knowledge routing to identify knowledge paths in FFNs and selectively update parameters during pre-training. Experimental results show that TRELM reduces pre-training time by at least 50% and outperforms other KEPLMs in knowledge probing tasks and multiple knowledge-aware language understanding tasks.

Unreliable feature extraction and matching in handcrafted features undermine the performance of visual SLAM in complex real-world scenarios. While learned local features, leveraging CNNs, demonstrate proficiency in capturing high-level information and excel in matching benchmarks, they encounter challenges in continuous motion scenes, resulting in poor generalization and impacting loop detection accuracy. To address these issues, we present DK-SLAM, a monocular visual SLAM system with adaptive deep local features. MAML optimizes the training of these features, and we introduce a coarse-to-fine feature tracking approach. Initially, a direct method approximates the relative pose between consecutive frames, followed by a feature matching method for refined pose estimation. To counter cumulative positioning errors, a novel online learning binary feature-based online loop closure module identifies loop nodes within a sequence. Experimental results underscore DK-SLAM's efficacy, outperforms representative SLAM solutions, such as ORB-SLAM3 on publicly available datasets., Comment: In submission

Recently, while large language models (LLMs) have demonstrated impressive results, they still suffer from hallucination, i.e., the generation of false information. Model editing is the task of fixing factual mistakes in LLMs; yet, most previous works treat it as a one-time task, paying little attention to ever-emerging mistakes generated by LLMs. We address the task of sequential model editing (SME) that aims to rectify mistakes continuously. A Dynamic Auxiliary Fusion Network (DAFNet) is designed to enhance the semantic interaction among the factual knowledge within the entire sequence, preventing catastrophic forgetting during the editing process of multiple knowledge triples. Specifically, (1) for semantic fusion within a relation triple, we aggregate the intra-editing attention flow into auto-regressive self-attention with token-level granularity in LLMs. We further leverage multi-layer diagonal inter-editing attention flow to update the weighted representations of the entire sequence-level granularity. (2) Considering that auxiliary parameters are required to store the knowledge for sequential editing, we construct a new dataset named \textbf{DAFSet}, fulfilling recent, popular, long-tail and robust properties to enhance the generality of sequential editing. Experiments show DAFNet significantly outperforms strong baselines in single-turn and sequential editing. The usage of DAFSet also consistently improves the performance of other auxiliary network-based methods in various scenarios, Comment: ACL2024 findings

Model editing aims to correct outdated or erroneous knowledge in large language models (LLMs) without the need for costly retraining. Lifelong model editing is the most challenging task that caters to the continuous editing requirements of LLMs. Prior works primarily focus on single or batch editing; nevertheless, these methods fall short in lifelong editing scenarios due to catastrophic knowledge forgetting and the degradation of model performance. Although retrieval-based methods alleviate these issues, they are impeded by slow and cumbersome processes of integrating the retrieved knowledge into the model. In this work, we introduce RECIPE, a RetriEval-augmented ContInuous Prompt lEarning method, to boost editing efficacy and inference efficiency in lifelong learning. RECIPE first converts knowledge statements into short and informative continuous prompts, prefixed to the LLM's input query embedding, to efficiently refine the response grounded on the knowledge. It further integrates the Knowledge Sentinel (KS) that acts as an intermediary to calculate a dynamic threshold, determining whether the retrieval repository contains relevant knowledge. Our retriever and prompt encoder are jointly trained to achieve editing properties, i.e., reliability, generality, and locality. In our experiments, RECIPE is assessed extensively across multiple LLMs and editing datasets, where it achieves superior editing performance. RECIPE also demonstrates its capability to maintain the overall performance of LLMs alongside showcasing fast editing and inference speed., Comment: 14 pages, 4 figures, 6 tables

Retrieval-augmented large language models (LLMs) leverage relevant content retrieved by information retrieval systems to generate correct responses, aiming to alleviate the hallucination problem. However, existing retriever-responder methods typically append relevant documents to the prompt of LLMs to perform text generation tasks without considering the interaction of fine-grained structural semantics between the retrieved documents and the LLMs. This issue is particularly important for accurate response generation as LLMs tend to ``lose in the middle'' when dealing with input prompts augmented with lengthy documents. In this work, we propose a new pipeline named ``Reinforced Retriever-Reorder-Responder'' (R$^4$) to learn document orderings for retrieval-augmented LLMs, thereby further enhancing their generation abilities while the large numbers of parameters of LLMs remain frozen. The reordering learning process is divided into two steps according to the quality of the generated responses: document order adjustment and document representation enhancement. Specifically, document order adjustment aims to organize retrieved document orderings into beginning, middle, and end positions based on graph attention learning, which maximizes the reinforced reward of response quality. Document representation enhancement further refines the representations of retrieved documents for responses of poor quality via document-level gradient adversarial learning. Extensive experiments demonstrate that our proposed pipeline achieves better factual question-answering performance on knowledge-intensive tasks compared to strong baselines across various public datasets. The source codes and trained models will be released upon paper acceptance.

With a fast developing pace of geographic applications, automatable and intelligent models are essential to be designed to handle the large volume of information. However, few researchers focus on geographic natural language processing, and there has never been a benchmark to build a unified standard. In this work, we propose a GeoGraphic Language Understanding Evaluation benchmark, named GeoGLUE. We collect data from open-released geographic resources and introduce six natural language understanding tasks, including geographic textual similarity on recall, geographic textual similarity on rerank, geographic elements tagging, geographic composition analysis, geographic where what cut, and geographic entity alignment. We also pro vide evaluation experiments and analysis of general baselines, indicating the effectiveness and significance of the GeoGLUE benchmark.

The minimal feature removal problem in the post-hoc explanation area aims to identify the minimal feature set (MFS). Prior studies using the greedy algorithm to calculate the minimal feature set lack the exploration of feature interactions under a monotonic assumption which cannot be satisfied in general scenarios. In order to address the above limitations, we propose a Cooperative Integrated Dynamic Refining method (CIDR) to efficiently discover minimal feature sets. Specifically, we design Cooperative Integrated Gradients (CIG) to detect interactions between features. By incorporating CIG and characteristics of the minimal feature set, we transform the minimal feature removal problem into a knapsack problem. Additionally, we devise an auxiliary Minimal Feature Refinement algorithm to determine the minimal feature set from numerous candidate sets. To the best of our knowledge, our work is the first to address the minimal feature removal problem in the field of natural language processing. Extensive experiments demonstrate that CIDR is capable of tracing representative minimal feature sets with improved interpretability across various models and datasets., Comment: Accepted by AAAI2024

Reasoning is a distinctive human capacity, enabling us to address complex problems by breaking them down into a series of manageable cognitive steps. Yet, complex logical reasoning is still cumbersome for language models. Based on the dual process theory in cognitive science, we are the first to unravel the cognitive reasoning abilities of language models. Our framework employs an iterative methodology to construct a Cognitive Tree (CogTree). The root node of this tree represents the initial query, while the leaf nodes consist of straightforward questions that can be answered directly. This construction involves two main components: the implicit extraction module (referred to as the intuitive system) and the explicit reasoning module (referred to as the reflective system). The intuitive system rapidly generates multiple responses by utilizing in-context examples, while the reflective system scores these responses using comparative learning. The scores guide the intuitive system in its subsequent generation step. Our experimental results on two popular and challenging reasoning tasks indicate that it is possible to achieve a performance level comparable to that of GPT-3.5 (with 175B parameters), using a significantly smaller language model that contains fewer parameters (<=7B) than 5% of GPT-3.5., Comment: emnlp 2023

Knowledge-Enhanced Pre-trained Language Models (KEPLMs) improve the performance of various downstream NLP tasks by injecting knowledge facts from large-scale Knowledge Graphs (KGs). However, existing methods for pre-training KEPLMs with relational triples are difficult to be adapted to close domains due to the lack of sufficient domain graph semantics. In this paper, we propose a Knowledge-enhanced lANGuAge Representation learning framework for various clOsed dOmains (KANGAROO) via capturing the implicit graph structure among the entities. Specifically, since the entity coverage rates of closed-domain KGs can be relatively low and may exhibit the global sparsity phenomenon for knowledge injection, we consider not only the shallow relational representations of triples but also the hyperbolic embeddings of deep hierarchical entity-class structures for effective knowledge fusion.Moreover, as two closed-domain entities under the same entity-class often have locally dense neighbor subgraphs counted by max point biconnected component, we further propose a data augmentation strategy based on contrastive learning over subgraphs to construct hard negative samples of higher quality. It makes the underlying KELPMs better distinguish the semantics of these neighboring entities to further complement the global semantic sparsity. In the experiments, we evaluate KANGAROO over various knowledge-aware and general NLP tasks in both full and few-shot learning settings, outperforming various KEPLM training paradigms performance in closed-domains significantly., Comment: emnlp 2023

Inspired by the discoveries in neuroscience, the method of visual place recognition develops toward using multiple homogenous spatial scales. We present a novel multi-scale place recognition algorithm mimicking the rodent map with multi-scale, discrete and overlapped characteristics. This visual system that can perform place recognition in the aerial environment without any constraint. We present a parallel and multi-channel processing network that can recognize places with a spatial scale and combine the output from these parallel processing channels. This recognizing network can utilize a multi-scale matching that builds associations between robotic activity and places at different spatial scales. Using two aerial datasets, the results demonstrate universal improvements achieved with multi-scale recognition approach. A systematic series of flight simulation experiments are conducted for analyzing the effect on the recognition and localization performance of varying matching scales. Finally, we present insights of further work in robotic navigation.

Accurately perceiving location and scene is crucial for autonomous driving and mobile robots. Recent advances in deep learning have made it possible to learn egomotion and depth from monocular images in a self-supervised manner, without requiring highly precise labels to train the networks. However, monocular vision methods suffer from a limitation known as scale-ambiguity, which restricts their application when absolute-scale is necessary. To address this, we propose SelfOdom, a self-supervised dual-network framework that can robustly and consistently learn and generate pose and depth estimates in global scale from monocular images. In particular, we introduce a novel coarse-to-fine training strategy that enables the metric scale to be recovered in a two-stage process. Furthermore, SelfOdom is flexible and can incorporate inertial data with images, which improves its robustness in challenging scenarios, using an attention-based fusion module. Our model excels in both normal and challenging lighting conditions, including difficult night scenes. Extensive experiments on public datasets have demonstrated that SelfOdom outperforms representative traditional and learning-based VO and VIO models., Comment: 14 pages, 8 figures, in submission

Recently, knowledge-enhanced pre-trained language models (KEPLMs) improve context-aware representations via learning from structured relations in knowledge graphs, and/or linguistic knowledge from syntactic or dependency analysis. Unlike English, there is a lack of high-performing open-source Chinese KEPLMs in the natural language processing (NLP) community to support various language understanding applications. In this paper, we revisit and advance the development of Chinese natural language understanding with a series of novel Chinese KEPLMs released in various parameter sizes, namely CKBERT (Chinese knowledge-enhanced BERT).Specifically, both relational and linguistic knowledge is effectively injected into CKBERT based on two novel pre-training tasks, i.e., linguistic-aware masked language modeling and contrastive multi-hop relation modeling. Based on the above two pre-training paradigms and our in-house implemented TorchAccelerator, we have pre-trained base (110M), large (345M) and huge (1.3B) versions of CKBERT efficiently on GPU clusters. Experiments demonstrate that CKBERT outperforms strong baselines for Chinese over various benchmark NLP tasks and in terms of different model sizes., Comment: EMNLP 2022 industry track

Chen Tian,1,* XiaoLei Xue,2,* Ye Chen,3,* Ruiyuan Liu,4 Yutong Wang,5 Sheng Ye,4 Zeyu Fu,5 Yingrui Luo,5 Shengmiao Wang,5 Xiaofeng He,1 Huajin Pang1 1Division of Vascular and Interventional Radiology, Department of General Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, Guangdong, 510515, People’s Republic of China; 2Department Pathology, Nanfang Hospital, Southern Medical University, Guangzhou, Guangdong, 510515, People’s Republic of China; 3Laboratory of Interventional Radiology, Department of Minimally Invasive Interventional Radiology and Department of Radiology, The Second Affiliated Hospital, Guangzhou Medical University, Guangzhou, Guangdong Province, 510260, People’s Republic of China; 4Guangdong Provincial Key Laboratory of Medical Image Processing, School of BioMedical Engineering, Southern Medical University, Guangzhou, Guangdong, 510515, People’s Republic of China; 5Cancer Research Institute, Experimental Education/Administration Center, School of Basic Medical Sciences, Southern Medical University, Guangzhou, Guangdong, 510515, People’s Republic of China*These authors contributed equally to this workCorrespondence: Xiaofeng He; Huajin Pang, Tel +86 13760661610, Email ozonetherapy@126.com; nfyyphj@163.comIntroduction: Second near-infrared photothermal therapy (NIR-II PTT) has become a promising strategy for treating cancer in terms of safety and potency. However, the application of NIR-II PTT was limited in the treatment of deep-buried solid tumors due to the low dose of NIR-II absorption nanomaterials and the inadequate laser energy in the deep tumor.Methods: Herein, the authors report the engineering of NIR-II absorbing polyaniline nanorods, termed HPW@PANI Nanorods, for in situ NIR-II PTT based on optical fibers transmission of laser power and transarterial infusion for the treatment of orthotopic hepatocellular carcinoma in the rabbit. HPW@PANI Nanorods were prepared via chemical ox

Distant supervision assumes that any sentence containing the same entity pairs reflects identical relationships. Previous works of distantly supervised relation extraction (DSRE) task generally focus on sentence-level or bag-level de-noising techniques independently, neglecting the explicit interaction with cross levels. In this paper, we propose a hierarchical contrastive learning Framework for Distantly Supervised relation extraction (HiCLRE) to reduce noisy sentences, which integrate the global structural information and local fine-grained interaction. Specifically, we propose a three-level hierarchical learning framework to interact with cross levels, generating the de-noising context-aware representations via adapting the existing multi-head self-attention, named Multi-Granularity Recontextualization. Meanwhile, pseudo positive samples are also provided in the specific level for contrastive learning via a dynamic gradient-based data augmentation strategy, named Dynamic Gradient Adversarial Perturbation. Experiments demonstrate that HiCLRE significantly outperforms strong baselines in various mainstream DSRE datasets.

Knowledge-Enhanced Pre-trained Language Models (KEPLMs) are pre-trained models with relation triples injecting from knowledge graphs to improve language understanding abilities. To guarantee effective knowledge injection, previous studies integrate models with knowledge encoders for representing knowledge retrieved from knowledge graphs. The operations for knowledge retrieval and encoding bring significant computational burdens, restricting the usage of such models in real-world applications that require high inference speed. In this paper, we propose a novel KEPLM named DKPLM that Decomposes Knowledge injection process of the Pre-trained Language Models in pre-training, fine-tuning and inference stages, which facilitates the applications of KEPLMs in real-world scenarios. Specifically, we first detect knowledge-aware long-tail entities as the target for knowledge injection, enhancing the KEPLMs' semantic understanding abilities and avoiding injecting redundant information. The embeddings of long-tail entities are replaced by "pseudo token representations" formed by relevant knowledge triples. We further design the relational knowledge decoding task for pre-training to force the models to truly understand the injected knowledge by relation triple reconstruction. Experiments show that our model outperforms other KEPLMs significantly over zero-shot knowledge probing tasks and multiple knowledge-aware language understanding tasks. We further show that DKPLM has a higher inference speed than other competing models due to the decomposing mechanism., Comment: Accepted by AAAI22

Recently, the performance of Pre-trained Language Models (PLMs) has been significantly improved by injecting knowledge facts to enhance their abilities of language understanding. For medical domains, the background knowledge sources are especially useful, due to the massive medical terms and their complicated relations are difficult to understand in text. In this work, we introduce SMedBERT, a medical PLM trained on large-scale medical corpora, incorporating deep structured semantic knowledge from neighbors of linked-entity.In SMedBERT, the mention-neighbor hybrid attention is proposed to learn heterogeneous-entity information, which infuses the semantic representations of entity types into the homogeneous neighboring entity structure. Apart from knowledge integration as external features, we propose to employ the neighbors of linked-entities in the knowledge graph as additional global contexts of text mentions, allowing them to communicate via shared neighbors, thus enrich their semantic representations. Experiments demonstrate that SMedBERT significantly outperforms strong baselines in various knowledge-intensive Chinese medical tasks. It also improves the performance of other tasks such as question answering, question matching and natural language inference., Comment: ACL2021

Precise geolocalization is crucial for unmanned aerial vehicles (UAVs). However, most current deployed UAVs rely on the global navigation satellite systems (GNSS) or high precision inertial navigation systems (INS) for geolocalization. In this paper, we propose to use a lightweight visual-inertial system with a 2D georeference map to obtain accurate and consecutive geodetic positions for UAVs. The proposed system firstly integrates a micro inertial measurement unit (MIMU) and a monocular camera as odometry to consecutively estimate the navigation states and reconstruct the 3D position of the observed visual features in the local world frame. To obtain the geolocation, the visual features tracked by the odometry are further registered to the 2D georeferenced map. While most conventional methods perform image-level aerial image registration, we propose to align the reconstructed points to the map points in the geodetic frame; this helps to filter out the large portion of outliers and decouples the negative effects from the horizontal angles. The registered points are then used to relocalize the vehicle in the geodetic frame. Finally, a pose graph is deployed to fuse the geolocation from the aerial image registration and the local navigation result from the visual-inertial odometry (VIO) to achieve consecutive and drift-free geolocalization performance. We have validated the proposed method by installing the sensors to a UAV body rigidly and have conducted two flights in different environments with unknown initials. The results show that the proposed method can achieve less than 4m position error in flight at 100m high and less than 9m position error in flight about 300m high.

Dynamic graph modeling has recently attracted much attention due to its extensive applications in many real-world scenarios, such as recommendation systems, financial transactions, and social networks. Although many works have been proposed for dynamic graph modeling in recent years, effective and scalable models are yet to be developed. In this paper, we propose a novel graph neural network approach, called TCL, which deals with the dynamically-evolving graph in a continuous-time fashion and enables effective dynamic node representation learning that captures both the temporal and topology information. Technically, our model contains three novel aspects. First, we generalize the vanilla Transformer to temporal graph learning scenarios and design a graph-topology-aware transformer. Secondly, on top of the proposed graph transformer, we introduce a two-stream encoder that separately extracts representations from temporal neighborhoods associated with the two interaction nodes and then utilizes a co-attentional transformer to model inter-dependencies at a semantic level. Lastly, we are inspired by the recently developed contrastive learning and propose to optimize our model by maximizing mutual information (MI) between the predictive representations of two future interaction nodes. Benefiting from this, our dynamic representations can preserve high-level (or global) semantics about interactions and thus is robust to noisy interactions. To the best of our knowledge, this is the first attempt to apply contrastive learning to representation learning on dynamic graphs. We evaluate our model on four benchmark datasets for interaction prediction and experiment results demonstrate the superiority of our model.

Reliably navigating to a distant goal remains a major challenge in robotics. In contrast, animals such as rats and pigeons can perform goal-directed navigation with great reliability. Evidence from neural science and ethology suggests that various species represent the spatial space as a topological template, with which they can actively evaluate future navigation uncertainty and plan reliable/safe paths to distant goals. While topological navigation models have been deployed in mobile robots, relatively little inspiration has drawn upon biology in terms of topological mapping and active path planning. In this paper, we propose a novel bio-inspired topological navigation model, which consists of topological map construction, active path planning and path execution, for aerial mobile robots with visual landmark recognition and compass orientation capability. To mimic the topological spatial representation, the model firstly builds the topological nodes based on the reliability of visual landmarks, and constructs the edges based on the compass accuracy. Then a reward diffusion algorithm akin to animals’ path evaluation process is developed. The diffusion process takes the topological structure and landmark reliability into consideration, which helps the agent to construct the path with visually reliable nodes. In the path execution process, the agent combines orientation guidance and landmark recognition to estimate its position. To evaluate the performance of the proposed navigation model, a systematic series of experiments were conducted in a range of challenging and varied real-world visual environments. The results show that the proposed model generates animal-like navigation behaviours, which avoids travelling across large visually aliased areas, such as forest and water regions, and achieves higher localization accuracy than navigating on the shortest paths.

Pre-trained neural language models bring significant improvement for various NLP tasks, by fine-tuning the models on task-specific training sets. During fine-tuning, the parameters are initialized from pre-trained models directly, which ignores how the learning process of similar NLP tasks in different domains is correlated and mutually reinforced. In this paper, we propose an effective learning procedure named Meta Fine-Tuning (MFT), served as a meta-learner to solve a group of similar NLP tasks for neural language models. Instead of simply multi-task training over all the datasets, MFT only learns from typical instances of various domains to acquire highly transferable knowledge. It further encourages the language model to encode domain-invariant representations by optimizing a series of novel domain corruption loss functions. After MFT, the model can be fine-tuned for each domain with better parameter initializations and higher generalization ability. We implement MFT upon BERT to solve several multi-domain text mining tasks. Experimental results confirm the effectiveness of MFT and its usefulness for few-shot learning., Comment: accepted by emnlp 2020

Lexical relations describe how concepts are semantically related, in the form of relation triples. The accurate prediction of lexical relations between concepts is challenging, due to the sparsity of patterns indicating the existence of such relations. We propose the Knowledge-Enriched Meta-Learning (KEML) framework to address the task of lexical relation classification. In KEML, the LKB-BERT (Lexical Knowledge Base-BERT) model is presented to learn concept representations from massive text corpora, with rich lexical knowledge injected by distant supervision. A probabilistic distribution of auxiliary tasks is defined to increase the model's ability to recognize different types of lexical relations. We further combine a meta-learning process over the auxiliary task distribution and supervised learning to train the neural lexical relation classifier. Experiments over multiple datasets show that KEML outperforms state-of-the-art methods., Comment: aaai 2021

In spite of the success of existing meta reinforcement learning methods, they still have difficulty in learning a meta policy effectively for RL problems with sparse reward. In this respect, we develop a novel meta reinforcement learning framework called Hyper-Meta RL(HMRL), for sparse reward RL problems. It is consisted with three modules including the cross-environment meta state embedding module which constructs a common meta state space to adapt to different environments; the meta state based environment-specific meta reward shaping which effectively extends the original sparse reward trajectory by cross-environmental knowledge complementarity and as a consequence the meta policy achieves better generalization and efficiency with the shaped meta reward. Experiments with sparse-reward environments show the superiority of HMRL on both transferability and policy learning efficiency., Comment: 13 pages

Machine Reading Comprehension (MRC) aims to extract answers to questions given a passage. It has been widely studied recently, especially in open domains. However, few efforts have been made on closed-domain MRC, mainly due to the lack of large-scale training data. In this paper, we introduce a multi-target MRC task for the medical domain, whose goal is to predict answers to medical questions and the corresponding support sentences from medical information sources simultaneously, in order to ensure the high reliability of medical knowledge serving. A high-quality dataset is manually constructed for the purpose, named Multi-task Chinese Medical MRC dataset (CMedMRC), with detailed analysis conducted. We further propose the Chinese medical BERT model for the task (CMedBERT), which fuses medical knowledge into pre-trained language models by the dynamic fusion mechanism of heterogeneous features and the multi-task learning strategy. Experiments show that CMedBERT consistently outperforms strong baselines by fusing context-aware and knowledge-aware token representations., Comment: ACL2021 (findings)

Efficient and robust visual place recognition is of great importance to autonomous mobile robots. Recent work has shown that features learned from convolutional neural networks achieve impressed performance with efficient feature size, where most of them are pooled or aggregated from a convolutional feature map. However, convolutional filters only capture the appearance of their perceptive fields, which lack the considerations on how to combine the multiscale appearance for place recognition. In this paper, we propose a novel method to build a multiscale feature pyramid and present two approaches to use the pyramid to augment the place recognition capability. The first approach fuses the pyramid to obtain a new feature map, which has an awareness of both the local and semi-global appearance, and the second approach learns an attention model from the feature pyramid to weight the spatial grids on the original feature map. Both approaches combine the multiscale features in the pyramid to suppress the confusing local features while tackling the problem in two different ways. Extensive experiments have been conducted on benchmark datasets with varying degrees of appearance and viewpoint variations. The results show that the proposed approaches achieve superior performance over the networks without the multiscale feature fusion and the multiscale attention components. Analyses on the performance of using different feature pyramids are also provided.

Graph representation learning, aiming to learn low-dimensional representations which capture the geometric dependencies between nodes in the original graph, has gained increasing popularity in a variety of graph analysis tasks, including node classification and link prediction. Existing representation learning methods based on graph neural networks and their variants rely on the aggregation of neighborhood information, which makes it sensitive to noises in the graph. In this paper, we propose Graph Denoising Policy Network (short for GDPNet) to learn robust representations from noisy graph data through reinforcement learning. GDPNet first selects signal neighborhoods for each node, and then aggregates the information from the selected neighborhoods to learn node representations for the down-stream tasks. Specifically, in the signal neighborhood selection phase, GDPNet optimizes the neighborhood for each target node by formulating the process of removing noisy neighborhoods as a Markov decision process and learning a policy with task-specific rewards received from the representation learning phase. In the representation learning phase, GDPNet aggregates features from signal neighbors to generate node representations for down-stream tasks, and provides task-specific rewards to the signal neighbor selection phase. These two phases are jointly trained to select optimal sets of neighbors for target nodes with maximum cumulative task-specific rewards, and to learn robust representations for nodes. Experimental results on node classification task demonstrate the effectiveness of GDNet, outperforming the state-of-the-art graph representation learning methods on several well-studied datasets. Additionally, GDPNet is mathematically equivalent to solving the submodular maximizing problem, which theoretically guarantees the best approximation to the optimal solution with GDPNet., Comment: 10 pages

Dynamic treatment recommendation systems based on large-scale electronic health records (EHRs) become a key to successfully improve practical clinical outcomes. Prior relevant studies recommend treatments either use supervised learning (e.g. matching the indicator signal which denotes doctor prescriptions), or reinforcement learning (e.g. maximizing evaluation signal which indicates cumulative reward from survival rates). However, none of these studies have considered to combine the benefits of supervised learning and reinforcement learning. In this paper, we propose Supervised Reinforcement Learning with Recurrent Neural Network (SRL-RNN), which fuses them into a synergistic learning framework. Specifically, SRL-RNN applies an off-policy actor-critic framework to handle complex relations among multiple medications, diseases and individual characteristics. The "actor" in the framework is adjusted by both the indicator signal and evaluation signal to ensure effective prescription and low mortality. RNN is further utilized to solve the Partially-Observed Markov Decision Process (POMDP) problem due to the lack of fully observed states in real world applications. Experiments on the publicly real-world dataset, i.e., MIMIC-3, illustrate that our model can reduce the estimated mortality, while providing promising accuracy in matching doctors' prescriptions., Comment: 10 pages, 11 figures. To appear in 2018 ACM SIGKDD Conference on Knowledge Discovery and Data Mining conference. Some typos are revise

Zelong Luo, Chao Feng, Peng Hu, Yong Chen, Xiao-feng He, Yanhao Li, Jianbo Zhao Department of Interventional Radiology, Nanfang Hospital, Southern Medical University, Guangzhou, Guangdong, People’s Republic of China Objective: The aim of this study was to investigate whether the level of serum microRNA-199a/b-3p (miR-199a/b-3p) can serve as a predictor of treatment response to transarterial chemoembolization (TACE) in patients with hepatocellular carcinoma (HCC).Methods: Serum miR-199a/b-3p expression level was measured in 132 patients with HCC before TACE (t1) and 3–5 days after TACE (t2). Additionally, 126 patients of these 132 patients had levels measured 4 weeks after TACE (t3) and 3–5 days after second TACE (t4). Serum miR-199a/b-3p expression levels were compared with those of 50 healthy controls. Correlations between miR-199a/b-3p expression levels and clinicopathologic factors and tumor responsiveness were analyzed. The modified Response Evaluation Criteria in Solid Tumors assessment was conducted at t3.Results: A lower mean baseline miR-199a/b-3p expression level was observed in patients with HCC compared with healthy controls (0.68±0.81 vs 2.50±2.16, P<0.001). A negative correlation between baseline miR-199a/b-3p expression levels and tumor size (P<0.001) was observed. The nonresponder group had significantly lower miR-199a/b-3p expression levels than the responder group at t1 (0.77±1.09 vs 1.96±1.32, P<0.001). In addition, the decrease in miR-199a/b-3p at t2 was greater in the responder group than in the nonresponder group (P=0.011). A higher proportion of the responder group achieved a >25% decrease in serum miR-199a/b-3p expression levels compared with the nonresponder group (64% vs 39%).Conclusion: Serum miR-199a/b-3p may represent a novel biomarker for predicting efficacy of TACE in patients with H

Titanium and its alloys have excellent properties and are promising biomaterial in medical engineering field. A bioactive surface on a Ti substrate is a prerequisite for great performance and long service life of implants. Based on the mechanism for inducing cell/tissue responses, three kinds of methods, namely morphological, physicochemical and biochemical methods, are reviewed in this paper. Hybrid methods that integrate individual methods or have additional functions are also discussed.

In order to control the high PV value between the vanes and the cam ring in a water hydraulic vane pump, a pressure-reducing valve with a constant pressure ratio was developed in this study. The pressure and leakage characteristics of the valve were theoretically analyzed by simulation and the experiments were also conducted based on the valve prototype. The theoretical analysis agrees well with the experimental results. Further, these results reveal that the pressure ratio of the valve decreases to the design value with the increasing working pressure. Additionally, the leakage of the valve rises simultaneously with the increasing working pressure and can be reduced significantly by decreasing the clearance between the valve body and the spool. The presented research not only proves the successful development of the pressure-reducing valve but also lays a foundation for the investigation of a high-pressure water hydraulic vane pump.

Micro hydroforming has an ability to manufacture complex 3D micro parts at a high production rate and has drawn increasing attentions. Brief understanding of macro hydroforming, for instance, deep drawing, is necessary to understand the principle of micro hydroforming. Then, special phenomena, such as size effects, occurred at micro scale are discussed and the related theories explaining these phenomena are introduced. Based on the similarities and differences between micro and macro hydroforming, experiments and simulation which consider the size effects are reviewed.

The design of a Fuzzy-PID controller involves setting the fuzzy rules, membership functions and its associated scaling factors. How to obtain a better control result and how these scaling factors affect the controller's performance are still a challenge. In this paper, the automatic position control system of a Hille 100 experimental rolling mill was used as a research testbed. Based on the mathematical control model of the rolling mill, a Fuzzy-PID controller was developed, and the process of implementing global optimization considering all these factors simultaneously by using genetic algorithm is introduced in detail. Through simulation, the performance of the control system with multifactor optimized Fuzzy-PID controller is given, and compared with that with only the fuzzy rules optimized in the controller. By simulation tests, it is found that these factors will influence the control performance of the controller, and that they are highly coupled with each other. The more factors for a Fuzzy-PID controller are optimized, the better the solution will be. It can also be inferred from the study that asymmetrical membership functions have more potential in improving a fuzzy controller's performance than symmetrical ones. The multifactor optimization method presented in this paper can in principle also be used to solve other complicated optimization issues.

The design of a Fuzzy-PID controller involves setting the fuzzy rules, membership functions and its associated scaling factors. How to obtain a better control result and how these scaling factors affect the controller's performance are still a challenge. In this paper, the automatic position control system of a Hille 100 experimental rolling mill was used as a research testbed. Based on the mathematical control model of the rolling mill, a Fuzzy-PID controller was developed, and the process of implementing global optimization considering all these factors simultaneously by using genetic algorithm is introduced in detail. Through simulation, the performance of the control system with multifactor optimized Fuzzy-PID controller is given, and compared with that with only the fuzzy rules optimized in the controller. By simulation tests, it is found that these factors will influence the control performance of the controller, and that they are highly coupled with each other. The more factors for a Fuzzy-PID controller are optimized, the better the solution will be. It can also be inferred from the study that asymmetrical membership functions have more potential in improving a fuzzy controller's performance than symmetrical ones. The multifactor optimization method presented in this paper can in principle also be used to solve other complicated optimization issues.

In this paper, finite element models of the strip shape during cold rolling of ultra thin strip in both symmetrical and asymmetrical rolling cases have been successfully developed, and the strip shape such as the thickness distribution along the strip width has been obtained. The strip shape and edge drop are discussed under both symmetrical and asymmetrical rolling conditions. Simulation results show that the asymmetrical rolling can reduce strip edge drop dramatically. The work roll edge curve also affects strip shape significantly. The developed finite element model has been verified with the experimental values.

Until now, most of the hydraulic hand tools for underwater application still use mineral oil as the system fluid medium. Due to the incompatibility with marine environment, these systems have highlighted a series of disadvantages. Therefore using seawater hydraulic tools has become a tendency in ocean engineering. In this paper, the design on a novel seawater hydraulic wirerope cutter for divers was introduced in detail based on its original prototype. A test rig with computer data acquisition system using Matlab/xPC was developed for testing the tool's performances. A group of wireropes with different diameters were used as specimen and they were cut off smoothly by the new developed tool in all the tests. The maximum pressures and necessary cutting forces for cutting off wirerope specimen were obtained by experiments. The results show that the cutting forces have linear relationship with the cross-sectional areas of the wireropes. It's verified by experiments that the tool is successfully designed and it has been applied in underwater application.

Until now, most of the hydraulic hand tools for underwater application still use mineral oil as the system fluid medium. Due to the incompatibility with marine environment, these systems have highlighted a series of disadvantages. Therefore using seawater hydraulic tools has become a tendency in ocean engineering. In this paper, the design on a novel seawater hydraulic wirerope cutter for divers was introduced in detail based on its original prototype. A test rig with computer data acquisition system using Matlab/xPC was developed for testing the tool's performances. A group of wireropes with different diameters were used as specimen and they were cut off smoothly by the new developed tool in all the tests. The maximum pressures and necessary cutting forces for cutting off wirerope specimen were obtained by experiments. The results show that the cutting forces have linear relationship with the cross-sectional areas of the wireropes. It's verified by experiments that the tool is successfully designed and it has been applied in underwater application.

Until now, most of the hydraulic hand tools for underwater application still use mineral oil as the system fluid medium. Due to the incompatibility with marine environment, these systems have highlighted a series of disadvantages. Therefore using seawater hydraulic tools has become a tendency in ocean engineering. In this paper, the design on a novel seawater hydraulic wirerope cutter for divers was introduced in detail based on its original prototype. A test rig with computer data acquisition system using Matlab/xPC was developed for testing the tool's performances. A group of wireropes with different diameters were used as specimen and they were cut off smoothly by the new developed tool in all the tests. The maximum pressures and necessary cutting forces for cutting off wirerope specimen were obtained by experiments. The results show that the cutting forces have linear relationship with the cross-sectional areas of the wireropes. It's verified by experiments that the tool is successfully designed and it has been applied in underwater application.

Until now, most of the hydraulic hand tools for underwater application still use mineral oil as the system fluid medium. Due to the incompatibility with marine environment, these systems have highlighted a series of disadvantages. Therefore using seawater hydraulic tools has become a tendency in ocean engineering. In this paper, the design on a novel seawater hydraulic wirerope cutter for divers was introduced in detail based on its original prototype. A test rig with computer data acquisition system using Matlab/xPC was developed for testing the tool's performances. A group of wireropes with different diameters were used as specimen and they were cut off smoothly by the new developed tool in all the tests. The maximum pressures and necessary cutting forces for cutting off wirerope specimen were obtained by experiments. The results show that the cutting forces have linear relationship with the cross-sectional areas of the wireropes. It's verified by experiments that the tool is successfully designed and it has been applied in underwater application.

A real-time control system for rolling mill control such as the Hille 100 experimental rolling mill was developed based on Matlab Simulink, Real-Time Workshop and xPC Target. The Host-Target PCs architecture for the control system was introduced, then the approach for the implementation of data acquisition, real-time control and the associated graphic user interface (GUI), were discussed in detail. Some limitations of the commercial software were also presented. The developed control system has been verified with Hille 100 rolling mill in laboratory, indicating the effective application in rolling mill control.

The demand of thin gauge strip with good quality such as the strip shape and surface finish is significantly increasing. In this study, finite element model of the strip shape during cold rolling of thin strip in asymmetrical rolling was developed, and the finite element simulation of the thin strip shape has been carried out in LS-DYNA. The effects of reduction and speed ratio on the strip shape and profile and the strip edge drop have been obtained. The developed finite element model has been verified with the experimental data. The obtained results are applicable to the control of the rolled thin strip shape in rolling practice.

The demand of thin gauge strip with good quality such as the strip shape and surface finish is significantly increasing. In this study, finite element model of the strip shape during cold rolling of thin strip in asymmetrical rolling was developed, and the finite element simulation of the thin strip shape has been carried out in LS-DYNA. The effects of reduction and speed ratio on the strip shape and profile and the strip edge drop have been obtained. The developed finite element model has been verified with the experimental data. The obtained results are applicable to the control of the rolled thin strip shape in rolling practice.