Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks454 from public data
- InfoXLM: An Information-Theoretic Framework for Cross-Lingual Language Model Pre-Training454
An information-theoretic framework that formulates cross-lingual language model pre- training as maximizing mutual information between multilingual-multi-granularity texts is presented and a new pre-training task based on contrastive learning is proposed.
- Bone Tissue Engineering in the Treatment of Bone Defects441
The 3D printing technology based on computer technology, aiming at designing personalized artificial scaffolds that can accurately fit bone defects, is summarized.
- Jointly Multiple Events Extraction via Attention-based Graph Information Aggregation413
This paper proposes a novel Jointly Multiple Events Extraction framework to jointly extract multiple event triggers and arguments by introducing syntactic shortcut arcs to enhance information flow and attention-based graph convolution networks to model graph information.
- 186
- Cross-Lingual Natural Language Generation via Pre-Training142
Experimental results on question generation and abstractive summarization show that the model outperforms the machine-translation-based pipeline methods for zero-shot cross-lingual generation and improves NLG performance of low-resource languages by leveraging rich-resource language data.
- XLM-E: Cross-lingual Language Model Pre-training via ELECTRA141
This paper introduces ELECTRA-style tasks to cross-lingual language model pre-training, and pretrain the model, named as XLM-E, on both multilingual and parallel corpora, which outperforms the baseline models on various cross-lingsual understanding tasks with much less computation cost.
- 123
- Read, Listen, and See: Leveraging Multimodal Information Helps Chinese Spell Checking118
The ReaLiSe model tackles the CSC task by capturing the semantic, phonetic and graphic information of the input characters, and selectively mixing the information in these modalities to predict the correct output.
- Deep Cross-Modal Hashing With Hashing Functions and Unified Hash Codes Jointly Learning99
A novel end-to-end Deep Cross-Modal Hashing with Hashing Functions and Unified Hash Codes Jointly Learning (DCHUC) is proposed, which outperforms the state-of-the-art cross-modal hashing methods.
- Improving Pretrained Cross-Lingual Language Models via Self-Labeled Word Alignment87
This paper introduces denoising word alignment as a new cross-lingual pre-training task that can serve as a pretrained word aligner, which achieves reasonably low error rate on the alignment benchmarks.
- Food recommendation with graph convolutional network83
This work proposes a novel model Food recommendation with Graph Convolutional Network (FGCN), which exploits ingredient-ingredient, ingredient-recipe, and recipe-user relations deeply and could alleviate the sparsity issue in food recommendation.
- mT6: Multilingual Pretrained Text-to-Text Transformer with Translation Pairs82
Experimental results show that the proposed mT6 improves cross-lingual transferability over mT5, and proposes a partially non-autoregressive objective for text-to-text pre-training.
- Social Influence Analysis: Models, Methods, and Evaluation71
This paper presents a survey of representative and state-of-the-art work in models, methods, and evaluation aspects related to SIA, and divides SIA models into microscopic and macroscopic models.
- Open Domain Event Extraction Using Neural Latent Variable Models67
This work considers open domain event extraction, the task of extracting unconstraint types of events from news clusters, and proposes a novel latent variable neural model, which is scalable to very large corpus.
- Rethinking Task-Oriented Dialogue Systems: From Complex Modularity to Zero-Shot Autonomous Agent62
This paper proposes a novel fully zero-013 shot autonomous TOD agent, named AutoTOD, where all the delicate modules in traditional TOD systems are deprecated and all it needs is a general-purpose instruction-following lan-017 guage model (e.g. GPT-4).
- Fundamental Capabilities of Large Language Models and their Applications in Domain Scenarios: A Survey62
A reliable strategy for domains to choose more robust backbone LLMs for real-world applications is proposed, establishing connections between fundamental capabilities and specific domains, and evaluating the varying importance of different capabilities.
- 58
- Cross-Lingual Abstractive Summarization with Limited Parallel Resources58
A novel Multi-Task framework for Cross-Lingual Abstractive Summarization (MCLAS) in a low-resource setting, Employing one unified decoder to generate the sequential concatenation of monolingual and cross-lingual summaries, MCLAS makes themonolingual summarization task a prerequisite of the CLS task.
- 54
- Earlier Attention? Aspect-Aware LSTM for Aspect-Based Sentiment Analysis54
A novel variant of L STM is proposed, termed as aspect-aware LSTM (AA-LSTM), which incorporates aspect information into LSTm cells in the context modeling stage before the attention mechanism, which can dynamically produce aspect- aware contextual representations.
- Task-oriented Word Embedding for Text Classification.54
A task-oriented word embedding method that regularizes the distribution of words to enable the embedding space to have a clear classification boundary and significantly outperforms the state-of-the-art methods.
- 53
- Query Expansion Based on a Feedback Concept Model for Microblog Retrieval53
This work proposes a Feedback Concept Model for query expansion that expands the query using knowledge information derived from Probase so that the expanded one could better reflect users' search intent, which allows for microblog retrieval at a concept-level, rather than term-level.
- Fundamental Capabilities and Applications of Large Language Models: A Survey50
This survey provides a systematic introduction to LLM’ fundamental capabilities, encompassing their definitions, formation mechanisms, and practical applications, and proposes evaluation strategies tailored to the essential capabilities required in each domain, offering practical guidance for selecting suitable backbone LLMs in real-world applications.
- 49
- CSE: Conceptual Sentence Embeddings based on Attention Model49
This work employs conceptualization model to assign associated concepts for each sentence in the text corpus, and then learns conceptual sentence embedding (CSE), which is more expressive than some widely-used text representation models such as latent topic model, especially for short-text.
- 48
- Neural Variational Correlated Topic Modeling47
This paper proposes a novel Centralized Transformation Flow to capture the correlations among topics by reshaping topic distributions and presents the Transformation Flow Lower Bound to improve the performance of the proposed model.
- Partial-Softmax Loss based Deep Hashing45
A novel Partial-Softmax Loss based Deep Hashing, called PSLDH, is proposed to generate high-quality hash codes and outperforms the state-of-the-art baselines in image retrieval task.
- A Supervised Multi-Head Self-Attention Network for Nested Named Entity Recognition43
This paper treats named entity recognition as a multi-class classification of word pairs and design a simple neural model, which can outperform the previous state-of-the-art methods on multiple tasks without any extra NLP tools or human annotations.
- BIT at SemEval-2017 Task 1: Using Semantic Information Space to Evaluate Semantic Textual Similarity42
This paper presents three systems for semantic textual similarity (STS) evaluation at SemEval-2017 STS task, which mainly depend on the (SIS), which is constructed based on the semantic hierarchical taxonomy in WordNet, to compute non-overlapping information content (IC) of sentences.
- 40
- Concept Pointer Network for Abstractive Summarization40
A concept pointer network that leverages knowledge-based, context-aware conceptualizations to derive an extended set of candidate concepts and points to the most appropriate choice using both the concept set and original source text.
- Cross-domain collaborative filtering with review text39
This paper extends previous transfer learning models in collaborative filtering, from linear mapping functions to non-linear ones, and proposes a cross-domain recommendation framework with the review text incorporated.
- Copy Is All You Need37
This paper formulate text generation as progressively copying text segments from an existing text collection, and compute the contextualized representations of meaningful text segments and index them using efficient vector search toolkits.
- Data-Aware Proxy Hashing for Cross-modal Retrieval37
A novel Data-Aware Proxy Hashing for cross-modal retrieval, called DAPH, which exploits the three types of data-aware proxy hash codes to supervise the training of modality-specific hashing networks and is able to generate discriminate hash codes with the semantic information preserved adequately.
- 37
- 35
- Camera Model Identification With Unknown Models35
The identification accuracy of SCIU is significantly better than the four state-of-art methods: 1) multi-class Support Vector Machine (SVM); 2) binary SVM; 3) combined classification framework; and 4) decision boundary carving.
- TemplateGEC: Improving Grammatical Error Correction with Detection Template34
A novel method, TemplateGEC, is proposed, which capitalizes on the capabilities of both Seq2Edit and Sequ2Seq frameworks in error detection and correction respectively, and reveals the potential of the method in performing human-in-the-loop GEC.
- 34
- Genre Separation Network with Adversarial Training for Cross-genre Relation Extraction34
A genre-separation network is designed, which applies two encoders, one genre-independent and onegenre-shared, to explicitly extract genre-specific and genre-agnostic features, and which outperforms the state-of-the-art by 1.7% absolute F1 gain.
- Neural machine translation with Gumbel Tree-LSTM based encoder33
This paper focuses on how to encode a source sentence into a vector in a unsupervised-tree way and then decode it into a target sentence and shows that the performance of the proposed model outperforms the attentional seq2seq model and the Transformer base model.
- Distant Supervision for Relation Extraction with Linear Attenuation Simulation and Non-IID Relevance Embedding33
This paper proposes a linear attenuation simulation to reflect the importance of words in the sentence with respect to the distances between entities and words, and a non-independent and identically distributed (non-IID) relevance embedding to capture the relevance of sentences in the bag.
- Discriminative Reasoning with Sparse Event Representation for Document-level Event-Event Relation Extraction32
This paper proposes a novel DECI model (SENDIR) for better document-level reasoning, different from existing works that build an event graph via linguistic tools, which introduces supportive event pairs for joint modeling.
- Improving neural machine translation with sentence alignment learning32
An NMT approach that heightens the adequacy in machine translation by transferring the semantic knowledge from bilingual sentence alignment learning by design a discriminator that learns to estimate sentence aligning score over translation candidates.
- 31
- Document-level relation extraction with Entity-Selection Attention30
This paper proposes a document-level relation extraction framework with two advantages that use the encoder to obtain the semantic features about the document and use the inter-sentence attention based on entity pairs to dynamically capture the features of multiple vital sentences.
- PONE30
This article proposes a novel learning-based metric that significantly improves the correlation with human judgments by using augmented POsitive samples and valuable NEgative samples, called PONE, and measures systematically all kinds of metrics to check which kind is best.
- 29
- A segment enhanced span-based model for nested named entity recognition29
The proposed Segment Enhanced Span-based model for nested NER has the advantages of enhancing boundary supervision in learning span representations by detecting segment endpoints, reducing the number of negative samples without losing long entities that are ignored by most span-based methods, and improving runtime performance.
- Supervised Deep Hashing for Hierarchical Labeled Data29
A novel deep hashing method, called supervised hierarchical deep hashing (SHDH), is proposed to perform hash code learning for hierarchical labeled data by weighting each level, and a deep neural network is designed to obtain a hash code for each data point.
- TCM-SD: A Benchmark for Probing Syndrome Differentiation via Natural Language Processing28
The core task of the TCM diagnosis and treatment system—syndrome differentiation (SD)—is focused on and the first public large-scale benchmark for SD is introduced, called TCM-SD, and a domain-specific pre-trained language model is proposed, called ZYBERT is proposed.
- Weighted Gaussian Loss based Hamming Hashing28
A novel Weighted Gaussian Loss based Hamming Hashing, called WGLHH, is proposed, which introduces a weighted Gaussian loss to optimize hashing model and outperforms the state-of-the-art baselines in image retrieval task.
- Domain-specific meta-embedding with latent semantic structures28
Experiments show that the proposed unsupervised meta-embedding method outputs more accurate word meta- embeddings for the task domain, compared to other state-of-the-art competitors.
- Low-resource Neural Machine Translation: Methods and Trends27
An overall review of existing deep learning techniques in low-resource NMT is made and the common characters among them are summarized and the future directions in this field are outlined.
- Sentence Semantic Matching Based on 3D CNN for Human–Robot Language Interaction26
A novel 3D CNN-based SSM (3DSSM) method for human–robot language interaction that is able to achieve comparable or even better performance over the state-of-the-art competing methods.
- 26
- Classifier-adaptation knowledge distillation framework for relation extraction and event detection with imbalanced data25
A Classifier-Adaptation Knowledge Distillation (CAKD) framework is proposed to address fundamental information extraction tasks, thus improving relation extraction and event detection performance and demonstrating the effectiveness of the proposed framework.
- 25
- Hypergraph network model for nested entity mention recognition24
A hypergraph network (HGN) model to recognize the nested entity mentions in texts that firstly uses encoders to extract the features and learn a hypergraph representation, and then recognizes entity mentions by tagging every hyperedge.
- 3-D Relation Network for visual relation recognition in videos23
A novel deep neural network is put forward, named 3DRN, to fuse the spatio-temporal visual characteristics, object label features, and spatial interactive features for learning the relation instances with multi-modal cues for visual relation recognition in videos.
- Penalized multiple distribution selection method for imbalanced data classification23
By proving that traditional classification methods that use single softmax distribution are limited for modeling complex and imbalanced data, a general Multiple Distribution Selection (MDS) method is proposed forImbalanced data classification, which outperforms previous methods.
- Semantic Structure-Based Word Embedding by Incorporating Concept Convergence and Word Divergence23
This paper argues that the structural information of well-organized words within the KBs is able to convey more effective and stable knowledge in capturing semantics of words, and introduces concept convergence and word divergence to reveal semantic structures in the word embedding learning process.
- Document-level Relation Extraction via Separate Relation Representation and Logical Reasoning22
This article proposes a novel model, SRLR, using Separate Relation Representation and Logical Reasoning considering the indirect relation representation and complex reasoning of evidence sentence problems, and proposes a logical reasoning module to identify evidence sentences and conduct relational reasoning.
- Neural abstractive summarization fusing by global generative topics22
This work proposes to incorporate a neural generative topic matrix as an abstractive level of topic information into a summarization generation system that is capable of generating succinct and recapitulative words or phrases.
- Conceptualization topic modeling22
The results show that the proposed models significantly outperform the baselines in terms of case study and perplexity, which means the new assumption is more reasonable than traditional one.
- Dimension reduction using collaborative representation reconstruction based projections22
This paper develops a collaborative representation reconstruction based projections (CRRP) method for dimension reduction that characterizes the between-class/within-class reconstruction error using collaborative representation and seeks the projections by maximizing thebetween- class reconstruction error to the within- Class reconstruction error.
- Enhancing camouflaged object detection through contrastive learning and data augmentation techniques21
- 21
- Exploring Dense Retrieval for Dialogue Response Selection21
This work presents a novel and efficient response selection model and a set of tailor-designed learning strategies to train it effectively, which notably outperforms the state-of-the-art baselines on both re-rank and full-rank evaluations.
- 21
- 21
- Steering Large Language Models for Cross-lingual Information Retrieval20
Activation Steered Multilingual Retrieval (ASMR) is presented that employs "steering activations''-a method to adjust and direct the LLM's focus-enhancing its ability to understand user queries and generate accurate, language-coherent responses.
- 20
- Unsupervised Hashing with Semantic Concept Mining20
A novel Unsupervised Hashing with Semantic Concept Mining, called UHSCM, which leverages a VLP model to construct a high-quality similarity matrix, and a novel hashing loss with a modified contrastive loss based regularization item is proposed to optimize the hashing network.
- Boosting legal case retrieval by query content selection with large language models20
Experimental results show that reformulating long queries using LLMs improves the performance of both sparse and dense models in legal case retrieval.
- From Question to Text20
A question-oriented feature attention (QFA) mechanism, which learns to weight different engineering features according to the given question, so that important features with respect to the specific question is emphasized accordingly.
- A P-LSTM Neural Network for Sentiment Classification20
A novel model based on long short-term memory recurrent neural network (L STM) called P-LSTM is proposed for sentiment classification, which achieves excellent performance on the sentiment classification tasks.
- 19
- GLEFFN: A Global-Local Event Feature Fusion Network for Micro-Expression Recognition19
This work marks the first successful attempt to solve the MER task from an event perspective, thus facilitating the future promotion of event-based camera technology and providing inspiration for future research endeavors in related domains.
- 19
- 19
- 18
- MindLLM: Pre-training Lightweight Large Language Model from Scratch, Evaluations and Domain Applications18
MindLLM is presented, a novel series of bilingual lightweight large language models, trained from scratch, alleviating such burdens by offering models with 1.3 billion and 3 billion parameters.
- 17
- Seeking Inspiration through Human-LLM Interaction16
The behavioral patterns of users actively seeking inspiration via LLMs are discussed, design opportunities for LLMs that make the inspiration-seeking process more human-centric are provided, and designs are provided that make the inspiration-seeking process more human-centric.
- Enlivening Redundant Heads in Multi-head Self-attention for Machine Translation16
This work proposes a redundant head enlivening (RHE) method to precisely identify redundant heads, and then vitalize their potential by learning syntactic relations and prior knowledge in the text without sacrificing the roles of important heads.
- Three birds, one stone: A novel translation based framework for joint entity and relation extraction16
- 16
- S2JSD-LSH: A Locality-Sensitive Hashing Schema for Probability Distributions16
This paper proposes a novel LSH scheme adapted to S2JSD-distance for approximate nearest neighbors search in high-dimensional probability distributions, and defines the specific hashing functions, and proves their local-sensitivity.
- I Know What You Want to Express: Sentence Element Inference by Incorporating External Knowledge Base16
- Lifetime Lexical Variation in Social Media15
A latent variable model that jointly models the lexical content of tweets and Twitter users’ ages and can learn meaningful age-specific topics such as “school” for teenagers and “health’ for older people is presented.
- How Users Interact with Generative Information Retrieval Systems: A Study of User Behavior and Search Experience14
A user study using Bing Chat to explore user behavior and feedback on generative IR systems and believes that this work provides valuable insight into the design and evaluation of generative information retrieval systems.
- 14
- A Hierarchical Context Augmentation Method to Improve Retrieval-Augmented LLMs on Scientific Papers14
This work proposes a hierarchical context augmentation method, which helps Retrieval-Augmented LLMs to autoregressively learn the structure knowledge of scientific papers, and utilizes the document tree to represent the hierarchical relationship of a paper and enhance the structure information of scientific context.
- 14
- 14
- Topic-related Chinese message sentiment analysis14
A graph-based ranking model of LT-IGT which takes both local and global topical information into consideration and a method of exploring sentimental features on expanded topical words with word embedding which considers both the syntactic and semantic information are proposed.
- Re-ranking voting-based answers by discarding user behavior biases14
This paper identifies the existence of two kinds of biases; position bias and appearance bias and proposes a joint click model for dealing with both of them and demonstrates how the ranking performance of the proposed model outperforms traditional methods with biases ignored.
- Emotional Tendency Identification for Micro-blog Topics Based on Multiple Characteristics14
This work constructs the dictionaries of sentiment words, internet slang and emoticons respectively, and implements the sentiment analysis algorithms based on phrase path and the multiple characteristics for emotional tendency of micro-blog topics to greatly improves the performance of emotional tendency identification on micro- blog topic.
- Feature Words Selection for Knowledge-based Word Sense Disambiguation with Syntactic Parsing14
With the help of syntactic parsing, the proposed methods can select feature words more accurately, which can alleviate the effect of noise words of window- based method and can avoid the paucity of feature words of dependency-based method.
- 14
- ALOHA: Adapting Local Spatio-Temporal Context to Enhance the Audio-Visual Semantic Segmentation13
The ALOHA framework, which Adapts LOcal spatio-temporal context to enHAnce AVSS and utilizes a parameter-shared encoder and applies the adapters in a staged manner to better adapt to embodied AI systems, significantly reduces the number of trainable parameters, making it more parameter-efficient.
- 13
- 13
- Can Monolingual Pretrained Models Help Cross-Lingual Classification?13
Two approaches to improve zero-shot cross-lingual classification, by transferring the knowledge from monolingual pretrained models to multilingual ones are presented.
- Combining External Sentiment Knowledge for Emotion Cause Detection13
A novel model to combine external sentiment knowledge for ECD task, namely ExSenti-ECD, that greatly outperforms the state-of-the-art baseline models and applies attention mechanism to take emotional context into consideration based on the observation that the context around emotion keywords can provide emotion cause clues.
- 13
- Forecasting Collector Road Speeds Under High Percentage of Missing Data13
This work proposes a multi-view road speed prediction framework that outperforms traditional approaches for arterial roads by 10% in MAE and RMSE, and in the Beijing road network, based on 10K taxi signals in 2 years.
- 12
- 12
- Learning Relation Ties with a Force-Directed Graph in Distant Supervised Relation Extraction12
A novel force-directed graph is proposed to comprehensively learn relation ties and is capable of modeling global correlation and mutual exclusion between relations, and outperforms the state-of-the-art baselines.
- Comprehensive Study: How the Context Information of Different Granularity Affects Dialogue State Tracking?12
This paper studies and discusses how the context information of different granularity affects dialogue state tracking and applies the findings about context granularity to few-shot learning scenario.
- SKR-QA: Semantic ranking and knowledge revise for multi-choice question answering12
It is argued that complete knowledge already exists in natural language and that the knowledge obtained by the Semantic-rank-and-Knowledge-Revise-based Question Answering approach is more conducive to machine understanding, thus providing certain interpretability.
- Weighted Aggregator for the Open-World Knowledge Graph Completion12
An aggregator is proposed, adopting an attention network to get the weights of words in the entity description, and this does not upset information in the word embedding, and make the single embedding of aggregation more efficient.
- Graph-based reasoning model for multiple relation extraction12
A corpus subgraph is proposed to mine more linguistic knowledge from the easily accessible unlabeled data, and sentence subgraphs are used to acquire classification knowledge, which jointly constitute a relation knowledge graph (RKG) to extract relations from sentences in this paper.
- Similarity-aware neural machine translation: reducing human translator efforts by leveraging high-potential sentences with translation memory12
This work firstly identifies testing sentences which are highly similar to training set (high-potential sentences) to reduce reviewing time, then focuses on improving corresponding translation quality greatly to reduce post-editing time.
- An Empirical Study of Classifier Combination Based Word Sense Disambiguation12
A new classifier combination based WSD method, namely the probability weighted voting method with dynamic self-adaptation, is proposed in this paper, which can take into consideration both the differences of classifiers and ambiguous instances.
- Towards Faithful Dialogues via Focus Learning11
Focus Learning (FocusL), a novel learning approach that adjusts the contribution of each token to the optimization direction by directly scaling the corresponding objective loss, is proposed.
- 11
- 11
- Community Detection Based on Minimum-Cut Graph Partitioning11
This paper proposes an efficient algorithm for community detection by adapting the Laplacian spectral partitioning algorithm and demonstrates that both the quality and the running time of the proposed algorithm rival the previous best algorithms.
- A Method of Polarity Computation of Chinese Sentiment Words Based on Gaussian Distribution11
A novel method of polarity computation of Chinese sentiment words based on Gaussian distribution is proposed which can analyze an analysis of semantic fuzziness ofChinese sentiment words quantitatively.
- Chinese Comparative Sentence Identification Based on the Combination of Rules and Statistics11
The method firstly normalizes the corpus and Chinese word segmentation, and then gets the broad extraction results by using comparative words, sentence structure templates and dependency relation analysis.
- 11
- 10
- Span-based Unified Named Entity Recognition Framework via Contrastive Learning10
A novel Span-based Unified NER framework via contrastive learning (SUNER), which aligns text span and entity type representations in a shared semantic space to extract entities in parallel, achieves better performance and higher efficiency than previous state-of-the-art unified NER models.
- Unifying Cross-lingual Summarization and Machine Translation with Compression Rate10
The proposed method not only improves the performance of the CLS task, but also provides controllability to generate summaries in desired lengths, and outperforms various strong baselines in three cross-lingual summarization datasets.
- Hashing based Efficient Inference for Image-Text Matching10
A novel hashing based efficient inference module called HEI is proposed, which can be plugged into the existing framework to speed up the inference step without reducing the retrieval performance.
- Domain structure-based transfer learning for cross-domain word representation10
This paper proposes a domain structure-based transfer learning method to learn cross-domain representations by leveraging the relations among domain-specific words, and constructs a semantic graph to capture the latent domain structure using domain- specific co-occurrence information.
- 10
- An input information enhanced model for relation extraction10
A novel end-to-end model to jointly extract semantic relations and argument entities from sentence texts, which achieves a significant improvement over the state-of-the-art model for relation extraction on the NYT dataset and achieves a competitive performance on the ACE2005 dataset.
- Sharing Pre-trained BERT Decoder for a Hybrid Summarization10
This paper proposes a novel extractive-and-abstractive hybrid framework for single document summarization task by jointly learning to select sentence and rewrite summary and applies the BERT pre-trained model as document encoder, sharing the context representations to both decoders.
- A Hybrid Method of Domain Lexicon Construction for Opinion Targets Extraction Using Syntax and Semantics10
A novel approach which incorporates domain lexicon with groups of syntactical and semantic features for opinion targets extraction of Chinese microblogs and Experimental results on COAE2014 dataset show the outperformance of the approach compared with other well-known methods.
- A noisy-smoothing relevance feedback method for content-based medical image retrieval10
A noisy-smoothing relevance feedback (NS-RF) method is proposed, which can deal with the noisy images caused by the users mislabeling some irrelevant images as relevant ones and a fuzzy support vector machine is adopted to re-rank the images.
- Micro-blog Post Topic Drift Detection Based on LDA Model10
A method of topic drift detection based on LDA model is proposed, using Gibbs sampling algorithm to obtain the probability distribution of micro-blog post words based on words correlation and detecting the topic drift by topic words sequence alignment based on discrete-time model.
- 10
- 9
- HSDS: An Abstractive Model for Automatic Survey Generation9
This paper proposes a novel abstractive method named Hierarchical Seq2seq model based on Dual Supervision (HSDS) to solve problems above, given multiple scientific papers in the same research area as input, the model aims to generate a corresponding survey.
- Query completion in community-based Question Answering search9
This work proposes a novel query completion algorithm based on ranking cQA questions with entity and phrase information for cZA search, and a demonstration system has been developed and shows that the proposed algorithm can successfully improve user experience.
- A novel unsupervised method for new word extraction9
A novel method to extract new words that not only considers domain specificity, but also combines with multiple statistical language knowledge to extract the top ranked new words.
- Social Media Processing9
This work proposes Tag Correspondence Model (TCM), a method to identify complex correspondences of tags from the rich context of microblog users to better understand the semantics of user tags.
- 8
- Interpretable modular knowledge reasoning for machine reading comprehension8
A modular knowledge reasoning approach based on neural network modules that explicitly model each reasoning process step, which leads to a more interpretable model and results in a more generalization capability.
- 8
- 8
- Self-supervised Bilingual Syntactic Alignment for Neural Machine Translation8
This work shows the first attempt of a source-target bilingual syntactic alignment approach SyntAligner by mutual information maximization-based self-supervised neural deep modeling, which firstly aligns the syntactic structures of source and target sentences and then maximizes their mutual dependency by introducing a lower bound on their mutual information.
- 8
- 8
- Labeled Phrase Latent Dirichlet Allocation and its online learning algorithm8
This paper proposes a novel topic model, called Labeled Phrase Latent Dirichlet Allocation (LPLDA), which regards each document as a mixture of phrases and partly considers the word order, and develops a batch inference algorithm based on Gibbs sampling technique for LPLDA.
- 8
- Collaborative representation analysis methods for feature extraction8
The experimental results on FERET and AR face databases show that CRA and KCRA are two effective feature extraction methods and could get good performance.
- Sentence Similarity Computational Model Based on Information Content8
Experimental results show that sentence similarity evaluated by the model is closer to human judgment than multiple competing baselines, and the proposed model neither needs any empirical parameter nor rely on other NLP tools.
- A Nonlinear Cross-Site Transfer Learning Approach for Recommender Systems8
This paper proposes a nonlinear transfer learning model, and utilizes the radial basis function (RBF) kernel to map user features of multiple sites, and demonstrates that the nonlinear property does exist in real recommender systems, which has been ignored previously.
- Graph Based Word Sense Disambiguation Method Using Distance Between Words8
Experimental results show that the proposed model can improve Chinese WSD performance, and compared with the best evaluation results of SemEval-2007: task #5, this model gets MacroAve(macro-average accuracy) increase 3.1%.
- An Improved Bat Algorithm with Doppler Effect for Stochastic Optimization8
Based on bats’ Doppler effect theory and the framework of the original bat algorithm, the new frequency equation in this algorithm is proposed, and the velocities and locations equation are also updated.
- An English-Chinese Cross-lingual Word Semantic Similarity Measure Exploring Attributes and Relations8
Experiments show that the proposed algorithm results in high correlation against human judgments, which encourages its broad application in cross-lingual applications.
- 8
- 7
- Online Sequential Decision-Making with Unknown Delays7
This work proposes three families of delayed algorithms based on approximate solutions to handle different types of received feedback and provides corresponding regret bounds under cases of general convexity and relative strong convexity.
- 7
- LPSA: Online educational video evaluation considering cognitive and teaching style7
The proposed LPSA video classification consisting of linguistic type (L), presentative type (P), scientific type (S), and artistic type (A), is proposed to realize teaching quality evaluation and learning process detection.
- 7
- Toward Real-life Dialogue State Tracking Involving Negative Feedback Utterances7
Simulation of negative feedback utterances in dialogue state tracking is explored in detail and three aspects of interaction mechanism that should be considered in real-life conversations involving negative Feedback utterances are explored and evaluated.
- Neural Variational Gaussian Mixture Topic Model7
A novel neural topic model called Neural Variational Gaussian Mixture Topic Model (NVGMTM) is proposed, which uses Gaussian distribution to depict the semantic relevance between words in the topics and outperforms the state-of-the-art baselines.
- 7
- 7
- Conceptual Multi-layer Neural Network Model for Headline Generation7
A new neural attention-based model called concept sensitive neural headline model is proposed, which connects the concept information of the document to input text for headline generation and achieves satisfactory results.
- Neural Chinese Word Segmentation as Sequence to Sequence Translation7
This paper casts the CWS as a sequence translation problem and proposes a novel sequence-to-sequence CWS model with an attention-based encoder-decoder framework, which captures the global information from the input and directly outputs the segmented sequence.
- 7
- Medical image retrieval based on unclean image bags7
A new three-stage solution to perform the retrieval and handle the noisy images, and a novel similarity aggregation method is proposed to incorporate image relevance and feature importance into the similarity computation process.
- Knowledge-based Word Sense Disambiguation with Feature Words Based on Dependency Relation and Syntax Tree7
A knowledgebased WSD method which obtains context feature words by dependency relation and syntax tree and the method is a preferred strategy to select feature words to disambiguate the target words.
- Survey of Chinese New Words Identification7
The status of the field for Chinese NWI was analyzed in detail, and the research techniques and existing problems for new words candidates extracting and new words POS guessing were discussed emphatically.
- 7
- DRMSpell: dynamically reweighting multimodality for Chinese spelling correction6
A multimodal pretrained language model called DRMSpell for CSC, which takes into consideration the interaction between the modalities, and a dynamically reweighting multimodality (DRM) module is introduced to reweight various modalities for obtaining more multimodal information.
- 6
- A Survey of Automatic Evaluation Methods on Text, Visual and Speech Generations6
A comprehensive review and a unified taxonomy of automatic evaluation methods for generated content across all three modalities is presented and five fundamental paradigms that characterize existing evaluation approaches across these domains are identified.
- AgentTOD: A Task-Oriented Dialogue Agent with a Flexible and Adaptive API Calling Paradigm6
This article proposes AgentTOD, a novel TOD framework that uses a large language model (LLM) as the intelligent agent to achieve a flexible dialogue paradigm that has superior performance on TODs as well as a superior adaptability to new task scenarios.
- Span Graph Transformer for Document-Level Named Entity Recognition6
This work proposes a novel Span Graph Transformer (SGT) method for document-level NER, which constructs long-range contextual dependencies at both the token and span levels and achieves better performance than previous state-of-the-art models.
- A syntactic evidence network model for fact verification6
The proposed SENet model has outperformed the baseline models and achieved the state-of-the-art performance for fact verification and the sentence attention mechanism is applied to obtain a richer semantic representation.
- 6
- 6
- 6
- 6
- Deep kernel supervised hashing for node classification in structural networks6
A novel Deep Kernel Supervised Hashing (DKSH) method to learn the hashing representations of nodes for node classification is proposed and significantly outperforms the state-of-the-art baselines over three real world benchmark datasets.
- A Finite-Time Consensus Using an Adaptive Terminal Sliding Mode Control Subjected to Input Saturation and Unknown Bounded Disturbance6
In this paper, finite-time consensus of double-integrator multi-agent systems is investigated and a new adaptive-terminal sliding mode control is proposed to satisfy the goal within a finite time by considering disturbances and input saturation.
- Towards Interpretable Reasoning over Paragraph Effects in Situation6
A sequential approach for the task of reasoning over paragraph effects in situation which explicitly models each step of the reasoning process with neural network modules, which leads to a more interpretable model.
- Multi-granularity semantic representation model for relation extraction6
A novel multi-granularity semantic representation (MGSR) model for relation extraction that can bridge the semantic gap between low-level semantic abstraction and high- level semantic abstraction by learning word-level, phrase- level, and sentence-level multi- granularity semantic representations successively.
- Exploring Explainable Selection to Control Abstractive Generation6
This paper targets using a select and generate paradigm to enhance the capability of selecting explainable contents and then guiding to control the abstract generation, and proposes a newly designed pair-wise extractor to capture the sentence pair interactions and their centrality.
- 6
- Preordering for Chinese-Vietnamese Statistical Machine Translation6
This paper has combined Chinese dependency relation and Chinese-Vietnamese word alignment results in order to pre-order Chinese word order to be suitable to Vietnamese one and shows that the methodology has improved the machine translation performance compared to the translation system using only the reordering models of phrase-based statistical machine translation.
- Tag Recommendation by Word-Level Tag Sequence Modeling6
This paper transforms tag recommendation into a word-based text generation problem and introduces a sequence-to-sequence model that inherits the advantages of LSTM-based encoder for sequential modeling and attention-based decoder with local positional encodings for learning relations globally.
- 6
- Locality-Sensitive Hashing for Finding Nearest Neighbors in Probability Distributions6
This paper presents a novel LSH scheme adapted to angular distance for ANN search in high-dimensional probability distributions, and proposes a Sequential Interleaving algorithm based on the “Unbalance Effect” of Euclidean and angular metrics for probability distributions.
- Query Intent Detection Based on Clustering of Phrase Embedding6
An efficient query intent generation model is proposed, which learns latent representations for intent candidates via embedding-based methods and significantly outperforms the state-of-art clustering algorithms in query intent detection.
- A Co-ranking Framework to Select Optimal Seed Set for Influence Maximization in Heterogeneous Network6
This paper forms the problem of influence maximization in heterogeneous network, and proposes a co-ranking framework to simultaneously select seed sets with different types, which is flexible and could adequately takes advantage of additional information implicit in the heterogeneous structure.
- When Factorization Meets Heterogeneous Latent Topics: An Interpretable Cross-Site Recommendation Framework6
This paper examines how to effectively and efficiently utilize cross-site ratings and content features to improve recommendation performance and how to make the recommendation interpretable by utilizing content features and proposes a joint model of matrix factorization and latent topic analysis.
- A study of age gaps between online friends6
A strong hypothesis based on homophily is validated and adapted to ensure the certainty of user attribute extracted via weakly supervised propagation and a propagation-based algorithm is designed to predict online users' age, leveraging the clique-based hypothesis.
- Unsupervised Word Sense Disambiguation Using Neighborhood Knowledge6
Experimental results indicate the proposed method can significantly outperform the baseline method and naturally make full use of the within-sentence relationship from the ambiguous sentence and cross-sentence relationship from the neighborhood knowledge.
- 5
- CXXC5 function blockade promotes diabetic wound healing through stimulating fibroblast and vascular endothelial cell activation5
Application of KY19382 improved these conditions, which promoted fibroblast activation and vascularization in high glucose environments and DM, which is an essential regulatory factor of wound healing and a prospective therapeutic target for treating chronic wound damage in diabetes.
- Electrospun 11β-HSD1 Inhibitor-Loaded Scaffolds for Accelerating Diabetic Ulcer Healing5
Evidence is provided for the use of 11β-HSD1 inhibitor-integrated biomaterials in treating DUs and represent a novel biological platform for modulating dysregulated mechanisms in DUs.
- 5
- Helping Language Models Learn More: Multi-Dimensional Task Prompt for Few-shot Tuning5
This work proposes MTPrompt, a multi-dimensional task prompt learning method consisting based on task-related object, summary, and task description information that achieves the best results on few-shot samples setting and five different datasets.
- Momentum Decoding: Open-ended Text Generation As Graph Exploration5
This study formulate open-ended text generation from a new perspective, i.e., it is viewed as an exploration process within a directed graph, and proposes a novel decoding method -- \textit{momentum decoding} -- which encourages the LM to explore new nodes outside the current graph.
- Unsupervised Question Answering via Answer Diversifying5
Extensive experiments show that the proposed method outperforms previous unsupervised models on five benchmark datasets, including SQuADv1.1, News QA, TriviaQA, BioASQ, and DuoRC, and shows strong performance in the few-shot learning setting.
- 5
- 5
- 5
- Generating Informative Dialogue Responses with Keywords-Guided Networks5
A simple but effective keywords-guided Sequence-to-Sequence model (KW-Seq2Seq) which uses keywords information as guidance to generate open-domain dialogue responses and produces more informative, coherent and fluent responses.
- Leveraging Hierarchical Deep Semantics to Classify Implicit Discourse Relations via a Mutual Learning Method5
A mutual learning neural model that makes use of multilevel semantic information together, including the distribution of implicit discourse relations, the semantics of arguments, and the co-occurrence of phrases and words is proposed.
- Quality Matters: Assessing cQA Pair Quality via Transductive Multi-View Learning5
A transductive multi-view learning model is presented, designed to find a latent common space by unifying and preserving information from various views, including question, answer, QA relevance, asker, and answerer, to enhance the representation ability of the common space.
- Labeled Phrase Latent Dirichlet Allocation5
A novel topic model, called Labeled Phrase LDA, which synchronously considers the supervised information and word order is proposed, which significantly outperforms baselines in terms of case study, perplexity and scalability.
- EHLLDA: A Supervised Hierarchical Topic Model5
A supervised hierarchical topic models are introduced: Extended Hierarchical Labeled Latent Dirichlet Allocation (EHLLDA), which aim to relax the assumptions of hLLDA by incorporating prior information of labels into hLLda.
- Named Entity Recognition Based on Bilingual Co-training5
This paper presents a semi-supervised approach to extract bilingual named entity, starting from a bilingual corpus where the named entities are extracted independently for each language, and a bilingual co-training algorithm is used to improve the named entity annotation quality.
- Active Learning for Cross Language Text Categorization5
This paper proposes an active learning algorithm for CLTC, which takes full advantage of both labeled data in the source language and unlabeledData in the target language, and extends the algorithm to double viewed form by considering the source and target language as two views of the classification problem.
- A Novel Spatial Clustering Analysis Method Using Bat Algorithm5
This paper presents a meta-modelling system that automates the very labor-intensive and therefore time-heavy and expensive and therefore expensive and expensive process of computer programming called “supervised learning”.
- 4
- 4
- Query in Your Tongue: Reinforce Large Language Models with Retrievers for Cross-lingual Search Generative Experience4
Built on the power of LLMs, MIMIR directly responds in the language of the user's query, reducing the need for post-search translations and effectively addressing traditional CLIR challenges.
- Beyond Labels and Topics: Discovering Causal Relationships in Neural Topic Modeling4
This paper proposes Causal Relationship-Aware Neural Topic Model (CRNTM), a novel neural topic model that can automatically uncover interpretable causal relationships between and within supervised information and latent topics, while concurrently discovering high-quality topics.
- 4
- Consprompt: Exploiting Contrastive Samples for Few-Shot Prompt Learning4
The proposed Consprompt combined with the prompt encoding network, contrastive sampling modules, and contrastive scoring modules, is introduced to realize differential contrastive learning.
- 4
- A syntactic multi-level interaction network for rumor detection4
A syntactic multi-level interaction network model which incorporates syntactic dependency relationships and multi-level interaction network for rumor detection and incorporates the extracted syntactic dependency relationships into the attention mechanism for language-driven word representation is proposed.
- A Passage-Level Reading Behavior Model for Mobile Search4
A probabilistic generative model, the Passage-level Reading behavior Model (PRM), is proposed, which outperforms existing unsupervised relevance estimation models and has strong interpretability and provides valuable insights into the understanding of how users seek and perceive useful information in mobile search.
- Reducing Length Bias in Scoring Neural Machine Translation via a Causal Inference Method4
This paper handles the length bias problem through a perspective of causal inference and applies a Half-Sibling Re-gression method to remove the length effect on S and then obtains a debiased translation score without length information.
- 4
- A Cross-Modal Classification Dataset on Social Network4
A novel Cross-Modal Classification Dataset constructed from Weibo called CMCD is constructed, which collects tweets with three modalities of text, image and video from 18 general categories, and filters tweets that can easily be classified by only textual contents.
- A Discriminative Convolutional Neural Network with Context-aware Attention4
This article proposes a discriminative CNN with context-aware attention to solve the challenges of vanilla CNN, and mainly encourages discrimination across different filters via maximizing their earth mover distances and estimates the salience of feature candidates by considering the relation between context features.
- Case-Sensitive Neural Machine Translation4
Two types of case-sensitive neural machine translation (NMT) approaches are introduced to alleviate the above problems: i) adding case tokens into the decoding sequence, and ii) adopting case prediction to the conventional NMT.
- Improving Neural Machine Translation by Achieving Knowledge Transfer with Sentence Alignment Learning4
This work designs a discriminator that learns to estimate sentence aligning score over translation candidates, and then the learned semantic knowledge is transfered to the NMT model under an adversarial learning framework and proposes a gated self-attention based encoder for sentence embedding.
- Multiple Perspective Answer Reranking for Multi-passage Reading Comprehension4
This study rearranges the order of traditional pipeline model and make a posterior answer reranking instead of prior passage reranking, and new proposed pre-trained language model BERT is also introduced here.
- Mapping sentences to concept transferred space for semantic textual similarity4
Experimental results suggest that the proposed comprehensive CTS-based model achieves significant improvement compared with the primitive one without the non-nominal words, OOV NEs, and word weights and also outperforms all the yearly state-of-the-art systems at the *SEM/SemEval 2013–2016 $$\mathcal {STS}$$STS tasks.
- QLUT at SemEval-2017 Task 2: Word Similarity Based on Word Embedding and Knowledge Base4
The results demonstrate that the combined method is effective for the computation of word similarity, while the word embeddings and the knowledge-based technique, respectively, needs more deeply improvement in details.
- From the perspective of teaching scheme4
The concrete requirements of "Output-Oriented" teaching evaluation in International Engineering Education Accreditation are analyzed, and a FCTS: Five-in-One Closed-loop Training Scheme is proposed, which is the key constituent part of innovation in BIT for cultivating engineering and technology professionals.
- Efficient Algorithm for Sentence Information Content Computing in Semantic Hierarchical Network4
An efficient method to calculate sentence information content, which employs the thinking of the difference set in hierarchical network is proposed, and experimental results show that the computational complexity decreases to O(n).
- A Novel Fast Framework for Topic Labeling Based on Similarity-preserved Hashing4
This paper proposes a novel fast topic labeling framework that casts the labeling problem as a k-nearest neighbor (KNN) search problem in a probability vector set, and shows that the proposed sequential interleaving method based on locality sensitive hashing (LSH) technology is efficient in boosting the comparison speed among probability distributions.
- BIT at SemEval-2016 Task 1: Sentence Similarity Based on Alignments and Vector with the Weight of Information Content4
Three unsupervised systems for determining the semantic similarity between two short texts or sentences submitted to the SemEval 2016 Task 1 are described, all of which make use of only off-the-shelf software and data making them easy to replicate.
- FCL: A New Network Words Extraction Approach Based on Statistical Language Knowledge4
This paper proposes a novel approach of FCL to extract new network words that not only considers domain specificity, but also combines with multiple statistical language knowledge.
- Chinese Word Sense Disambiguation based on Context Expansion4
Experimental results show that the second method can significantly improve traditional WSD accuracy by 2.21%.
- Incorporating New Words Detection with Chinese Word Segmentation4
The author reviewed the problem caused by the new words in Chinese texts, then the algorithm of new words detection was introduced, and the official evaluation result in this bakeoff was provided.
- Analyzing Fair Exchange Protocols with Strand Spaces4
A recursive analysis method is constructed according to the idea of message drive of fair exchange protocols, and a ZG fair non-repudiation protocol is analyzed and found a flaw which wasn't detected by some other analysis.
- 3
- 3
- 3
- Laser: Parameter-Efficient LLM Bi-Tuning for Sequential Recommendation with Collaborative Information3
A parameter-efficient Large Language Model Bi-Tuning framework for sequential recommendation with collaborative information (Laser), and M-Former, a lightweight MoE-based querying transformer that uses a set of query experts to integrate diverse user-specific collaborative information encoded by frozen ID-based sequential recommender systems, significantly improving the accuracy of recommendations.
- Optimizing Chain-of-Thought Reasoning: Tackling Arranging Bottleneck via Plan Augmentation3
This study subdivide CoT reasoning into two parts: arranging and executing, and identifies that the bottleneck of models mainly lies in arranging rather than executing, and proposes a plan-based training and reasoning method that guides models to generate arranging steps through abstract plans.
- Can Cross-Lingual Transferability of Multilingual Transformers Be Activated Without End-Task Data?3
Experimental results show that P LUG I N -X successfully activates the cross-lingual transferability of multilingual Transformers without accessing end-task data.
- 3
- 3
- 3
- ET5: A Novel End-to-end Framework for Conversational Machine Reading Comprehension3
A novel end-to-end framework for conversational machine reading comprehension based on shared parameter mechanism, called entailment reasoning T5 (ET5), is proposed, which achieves new state-of-the-art results on the ShARC leaderboard with the BLEU-4 score of 55.2.
- Hammer PDF: An Intelligent PDF Reader for Scientific Papers3
The proposed Hammer PDF Reader can help researchers, especially those studying computer science, to improve the efficiency and experience of reading scientific papers.
- Location-based Twitter Filtering for the Creation of Low-Resource Language Datasets in Indonesian Local Languages3
This paper identifies the problems faced when constructing a Local Indonesian NLP dataset and is developing a framework for creating, collecting, and classifying Local Indonesian datasets for NLP.
- 3
- 3
- 3
- Multi-task Learning for Low-Resource Second Language Acquisition Modeling3
This paper proposes a novel SLA modeling method, which learns the latent common patterns among different language-learning datasets by multi-task learning and are further applied to improving the prediction performance in low-resource scenarios.
- 3
- Integrating pronunciation into Chinese-Vietnamese statistical machine translation3
It is found that the pronunciation features can improve both Chinese-Vietnamese and Vietnamese translation qualities, and the translation performance is improved by a maximum value of 1.03.
- Visually-aware Collaborative Food Recommendation.3
A dedicated neural network-based solution Hierarchical Attention based Food Recommendation (HAFR) which is capable of capturing the collaborative filtering effect like what similar users tend to eat and inferring a user's preference at the ingredient level and learning user preference from the recipe's visual image is developed.
- Collective Entity Linking on Relational Graph Model with Mentions3
A collective entity linking model which effectively leverages the global interdependence among mentions in the same source document and unifies semantic relations and co-reference relations into relational inference for semantic information extraction is proposed.
- Supervised Hashing for Multi-labeled Data with Order-Preserving Feature3
This work proposed a novel supervised hashing method to perform hash codes learning for multi-labeled data that extracts the order-preserving data features through deep convolutional neural network and would be used for learning hash codes.
- Singleton Detection for Coreference Resolution via Multi-window and Multi-filter CNN3
A novel singleton detection method based on multi-window and multi-filter convolutional neural network (MMCNN) that can detect singleton mentions with less use of hand-designed features and more sentence information is proposed.
- Aligning Gaussian-Topic with Embedding Network for Summarization Ranking3
The proposed model incorporates the topical aspects and continuous vector representations, which jointly learns semantic rich representations encoded by vectors, and is leveraged by topic filtering and embedding ranking model, which can select desirable salient sentences.
- Utilizing Crowdsourcing for the Construction of Chinese-Mongolian Speech Corpus with Evaluation Mechanism3
This paper explores the use of Wechat Official Account Platform (WOAP) in order to build a speech corpus and to assess the feasibility of using WOAP followers (also known as contributors) to assemble speech corpus of Mongolian.
- 3
- 3
- An Online Inference Algorithm for Labeled Latent Dirichlet Allocation3
This paper develops an efficient online algorithm for Labeled LDA, called online-LLDA, based on particle filter, a Sequential Monte Carlo approximation technique, which significantly outperforms batch algorithm (batch- LLDA) in time, while preserving equivalent quality.
- 3
- 3
- Distinguishing Social Ties in Recommender Systems by Graph-Based Algorithms3
This paper proposes to employ three graph-based algorithms to distinguish and propagate the influence among the friends of an active user, and then integrate them into the factorization-based social recommendation framework.
- A Survey on Chinese Chunk Parsing3
This paper surveys the rich researches on chunking in several aspects: the definition and classification of chunks, the chunk identification, the chunks annotation and evaluation, and the internal relationship in chunks.
- Substring reduction algorithm based on independence statistic3
According to the analysis of the one to multiple relationship between the meaningless substring and its parent strings, a substring reduction algorithm based on independence statistic is proposed to filter these meaningless substrings.
- 3
- 2
- 2
- FlashBack: Efficient Retrieval-Augmented Language Modeling for Fast Inference2
FlashBack is proposed, a modular RALM designed to improve the inference efficiency of RALM with appending context pattern while maintaining decent performance after fine-tuning by Low-Rank Adaption and introduces Marking Token as two special prompt tokens for marking the boundary of the appending context during fine-tuning.
- 2
- 2
- From Skeleton to Flesh: Aggregated Relational Transformer Towards Controllable Video Captioning with Two-Step Decoding2
This work introduces the review process in the Transformer-based framework and proposes a novel network, Aggregated Relational Transformer (ART) to conduct two-step decoding for video captioning, capable of controlling captions to respond to different semantic contexts.
- 2
- 2
- 2
- 2
- Which is more faithful, seeing or saying? Multimodal sarcasm detection exploiting contrasting sentiment knowledge2
A contrasting sentiment‐based model for multimodal sarcasm detection (CS4MSD), which identifies inconsistent emotions by leveraging the CLIP knowledge module to produce sentiment features in both text and image.
- 2
- 2
- Towards Efficient Coarse-grained Dialogue Response Selection2
This work proposes an efficient coarse-grained response selection subsystem consisting of two novel methods, the Hierarchical Dense Retrieval and the Deep Semantic Hashing, which achieves over 5x FLOPs speedup and over 192x storage compression ratio.
- 2
- 2
- 2
- 2
- 2
- Piecewise graph convolutional network with edge-level attention for relation extraction2
A novel model, PGCN-EA, using Piecewise Graph Convolutional Network with Edge-level Attention to address redundant syntactic information and wrong dependency parsing results is proposed, and the piecewise adjacency matrix based on entity pair is employed.
- Building Knowledge-Grounded Dialogue Systems with Graph-Based Semantic Modeling2
A novel graph structure is presented, Ground Graph, which models the semantic structure of both dialogue contexts and knowledge documents to facilitate knowledge selection and integration for the task and shows excellent generalization ability in resource-limited situations.
- A Prototype-Based Few-Shot Named Entity Recognition2
This work proposes the ClusLoss and the ProEuroLoss aiming to enhance the model's ability in terms of aggregating semantic information spatially, thus helping the model better distinguish entity types.
- 2
- 2
- 2
- Improving Neural Machine Translation by Transferring Knowledge from Syntactic Constituent Alignment Learning2
This article proposes to improve NMT by explicitly learning the bilingual syntactic constituent alignments, and utilizes syntactic parsers to induce syntactic structures of sentences, and proposes two ways to utilize the syntactic constituents in a perceptual (not adversarial) generator-discriminator training framework.
- 2
- 2
- 2
- 2
- SciNER: A Novel Scientific Named Entity Recognizing Framework2
A novel scientific named entity recognizing pipeline framework, called SciNER, which outperforms state-of-the-art baselines on several datasets in both computer science and biomedical domains.
- 2
- Fast and Accurate Bilingual Lexicon Induction via Matching Optimization2
Empirical experiments demonstrate that the proposed BLIMO method outperforms state-of-the-arts baselines greatly on two standard benchmarks and converts the lexicon induction problem into a maximum weighted matching problem, which could be efficiently solved by the matching optimization with greedy search.
- 2
- 2
- Training Set Similarity Based Parameter Selection for Statistical Machine Translation2
Experimental results show that the proposed model parameters selection method leads to the improvements up to 2.4 and 2.2 BLEU score respectively, and it also shows the good interpretability of the proposed method.
- 2
- Leveraging Pattern Associations for Word Embedding Models2
This paper proposes a novel approach to combine associated patterns for word embedding method via joint training objection and shows that the proposed method can perform significantly better than the state-of-the-arts baseline models.
- Identifying Suspected Cybermob on Tieba2
This paper focuses on the origin of information diffusion, namely back to its makers and contributors, and constructs real system to identify suspected cybermob automatically using machine learning method with above features, including other common features like user/content-based ones.
- Conceptual Sentence Embeddings2
This work employs concept conceptualization model to assign associated concepts for each sentence in the text corpus, and learns conceptual sentence embedding (CSE), which is more expressive than some widely-used document representation models such as latent topic models.
- Opinion Mining Research on Chinese Micro-blog2
This work proposes to simultaneously identify opinion targets and opinion-bearing words based on multi-features in Chinese micro-blog post and decomposes the task into three phases.
- Patent Literatures Translation System Based on Hadoop2
The experimental results show that the proposed machine translation system in this paper has better translation performance than the conventional machine translation approach.
- 2
- Researches on Ontology-based Technical Lexicons for Specialty Machine Translation2
Some important aspects of specialty MT-oriented technical lexicons are studied, including the design of general-purpose specialty ontology, the description of technical terms and their mapping to specialist ontology and the organization and application of bilingual or multilingual MT domain-specific lexicons.
- EmergentBridge: Improving Zero-Shot Cross-Modal Transfer in Unified Multimodal Embedding Models1
EmergentBridge is proposed, an embedding-level bridging framework that improves performance on unpaired pairs without requiring exhaustive pairwise supervision, and consistently outperforms prior binding baselines on zero-shot classification and retrieval.
- Look, Listen and Segment: Towards Weakly Supervised Audio-Visual Semantic Segmentation1
Progressive Cross-modal Alignment for Semantics (PCAS) builds a classification task to train the audio-visual encoder using video labels, injects visual semantic prompts to enhance frame-level audio understanding, and then applies progressive contrastive alignment to map audio categories to image regions without mask annotations.
- 1
- EXCEEDS: Extracting Complex Events via Nugget-based Grid Modeling in Scientific Domain1
This work proposes EXCEEDS, an end-to-end scientific event extraction framework by encoding dense nuggets into a grid matrix and simplifying complex event extraction as a nugget-based grid modeling task.
- 1
- 1
- 1
- MMWOZ: Building Multimodal Agent for Task-oriented Dialogue1
A novel multimodal model called MATE (Multimodal Agent for Task-oriEnted dialogue) is proposed as the baseline model for the MMWOZ dataset and comprehensive experimental analysis is conducted using MATE to investigate the construction of a practical multimodal agent for task-oriented dialogue.
- ESTJ: Enhancing Structured Tendency Judgment in Hybrid-Modal Table Understanding1
A modality preference alignment model, called ESTJ, is proposed, called ESTJ, which Enhances Structured Tendency Judgment in HMTU, which alleviates modality preference imbalance by adhering to the principle of least modality priority.
- 1
- 1
- 1
- Alleviating repetitive tokens in non-autoregressive machine translation with unlikelihood training1
This study proposes a novel approach to train a non-autoregressive model using unlikelihood loss and demonstrates that this approach significantly reduces the number of repetitive tokens while improving the overall performance of non-autoregressive machine translations.
- 1
- 1
- 1
- 1
- Approximating to the Real Translation Quality for Neural Machine Translation via Causal Motivated Methods1
This article studies the length bias problem of beam search for neural machine translation (NMT) and the input related noise problem of translation quality estimation (QE) and demonstrates the rationality of the empirical assumptions made on the methods.
- CPLR-SFS: Contrastive Prompt Learning to Reduce Redundancy for Scientific Faceted Summarization1
This work proposes a novel Contrastive Prompt Learning method to Reduce redundancy for Scientific Faceted Summarization, named CPLR-SFS, to generate concise and less-overlapping faceted summaries.
- SNER-CS: Self-training Named Entity Recognition in Computer Science1
A novel self-training method based pretraining language model with a distantly supervised label automatic construction system in CS (SNER-CS) is proposed, which performs previous state-of-the-art methods in computer science NER task.
- Graph vs. Sequence: An Empirical Study on Knowledge Forms for Knowledge-Grounded Dialogue1
This work conducts a thorough experiment and study on the task of generating an informative response based on both the dialogue history and external knowledge source, and offers conclusive solutions and sensible suggestions for directions and standards of future research.
- Better Localness for Non-Autoregressive Transformer1
This work designs a dynamic mask matrix according to the query tokens, key tokens, and relative distance, and unify the localness module for self-attention and cross-att attention module and shows that this model can significantly improve the performance of the non-autoregressive transformer.
- SciMRC: Multi-perspective Scientific Machine Reading Comprehension1
A novel multi-perspective scientific machine reading comprehension dataset, SciMRC, which incorporates perspectives from beginners, students, and experts is introduced, which highlights the challenging nature of the scientific machine comprehension tasks.
- 1
- Improving Non-Autoregressive Machine Translation Using Sentence-Level Semantic Agreement1
A new method is proposed, Sentence-Level Semantic Agreement (SLSA), to obtain consistency between the source and generated sentences, and the improvement in the NAT model’s performance shows the effect of the proposed method.
- 1
- 1
- 1
- Enhanced encoder for non-autoregressive machine translation1
Experimental results on three translation tasks show that the proposed approach to enhance the encoder’s modeling ability by using a pre-trained BERT model as an extra encoder can significantly improve the performance of non-autoregressive MT, and surpass the baseline non- autoregressive models.
- 1
- 1
- A Span-Based Distantly Supervised NER with Self-learning1
A novel span-based self-learning method, which employs span-level features to update corresponding dictionaries based on both local and global features, and performs better than the state-of-the-art baselines.
- 1
- Streamlined Decoder for Chinese Spoken Language Understanding1
A novel streamlined decoding structure with attention mechanism, which uses three successively connected RNN to decode act, slot and value respectively, which exceeds state-of-the-art model on an unaligned multi-turn task-oriented Chinese spoken dialogue dataset provided by the contest.
- 1
- The construction and practice of integration cultivation mechanism for innovative talents in CS discipline1
ICM4IT from the five aspects in details is introduced, idea (Outcomes-Based Education, OBE), mechanism (Bachelor-Master-PhD penetration cultivation), mode (Multi-layer/mode/platform internationalization), practice (MOOC/SPOC-based online & offline teaching), support (research & education coordination).
- Zewen at SemEval-2018 Task 1: An Ensemble Model for Affect Prediction in Tweets1
This paper presents a method for Affect in Tweets, which is the task to automatically determine the intensity of emotions and intensity of sentiment of tweets, and proposes an en-semble system including four different deep learning methods which are CNN, Bidirectional LSTM (BLSTM), L STM-CNN and a CNN-based Attention model.
- Assembling Chinese-Mongolian Speech Corpus via Crowdsourcing1
It is shown that WeChat Official Account Platform can be used to assemble speech corpus quickly and cheaply, with near-expert accuracy, and the construction of bilingual speech corpus via crowdsourcing has a reference value for the similar studies.
- Incorporating target language semantic roles into a string-to-tree translation model1
Two methods to use semantic roles to improve the performance of the string-to-tree translation model by adding role labels in the syntax tree and constructing a semantic role tree, and incorporating the syntax information into it are proposed.
- Leveraging Hierarchical Deep Semantics to Classify Implicit Discourse Relations via Mutual Learning Method1
The results of both binary and multiclass identification show that this method outperforms previous works since the mutual learning strategy can distinguish Expansion type from the others efficiently.
- A Hybrid Sentence Splitting Method by Comma Insertion for Machine Translation with CRF1
A log linear model is used to insert commas into proper positions to split long sentence, trying to shorten the length of sub-sentence and benefit to machine translation.
- Emotional Element Extraction Based on CRFs1
This paper designs several rule-based methods to extract evaluation phrase which is in complex structure, as well as comment targets and opinion holders, and improves the performance of emotional element extraction in the domain of sentiment analysis for automobile's Chinese comments.
- 1
- 1
- Effectiveness Comparison of Feature Words for Word Sense Disambiguation1
The results of experiments show that the method to select feature words and assign their weights with dependency tree is a preferred strategy for WSD.
- Do ads compete or collaborate?1
A novel click-model, Full Relation Model (FRM), is proposed, which explicitly models dependencies between ads, and shows significant improvement in CTR prediction as compared to earlier click models.
- 1
- Chinese Personal Name Disambiguation Based on Person Modeling1
The authors introduce the frame of person disambiguation system LJPD, which uses a new person model, which uses a new person model and shows that the precision is competitive, but the recall is very low.
- 1
- Design and implementation of FAQ automatic return system based on similarity computation1
This paper introduces the design and implementation of a FAQ automatic return system based on semantic similarity computation, including computation model choosing, FAQ characters analyzing, FAQ data formal expressing, feature vector indexing, and weight computing and so on.
- –
- Planner Policies Are Bias Learners: A Comprehensive Analysis of Preference Bias in LLM-based Dialogue Planner Policies–
This article systematically evaluates four state-of-the-art planning methods across three dialogue domains, utilizing multiple LLM families at varying scales, and reveals that all tested planners exhibit significant preference bias.
- MDEFusion: A Multi‐Domain EEG Feature Fusion Network With Bidirectional Attention LSTM and Half‐Split Crossover SAE for Schizophrenia Recognition–
Using electroencephalogram signals for objective and accurate recognition of SZ is critical for timely clinical intervention and existing recognition methods still face key challenges, including incomplete feature extraction, limited computational efficiency, and insufficient modeling of long‐range temporal dependencies.
- –
- –
- –
- Training-free Truthfulness Detection via Sparse MLP Value Vectors–
This work proposes TruthV, a training-free approach that operates at the level of individual multi-layer perceptron (MLP) value vectors, demonstrating that truthfulness-related variation in LLMs is captured in a sparse and structured manner at the level of MLP value vectors.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- LLMs Judge Themselves: A Game-Theoretic Framework for Human-Aligned Evaluation–
This work proposes a novel alternative to conventional evaluation practices, where LLMs assess each other's output through self-play and peer review, and incorporates game-theoretic voting algorithms to aggregate peer reviews, enabling a principled investigation into whether model-generated rankings reflect human preferences.
- –
- A Distributed Collaborative Retrieval Framework Excelling in All Queries and Corpora based on Zero-shot Rank-Oriented Automatic Evaluation–
A novel Distributed Collaborative Retrieval Framework is proposed, outperforming each single model across all queries and corpora, and can achieve performance comparable to effective listwise methods like RankGPT and ListT5, while offering superior efficiency.
- Identifying and Analyzing Performance-Critical Tokens in Large Language Models–
This work provides additional insights into how large language models (LLMs) leverage task reasoning procedures in ICL and suggests that future work may involve using task-encoding tokens to improve the computational efficiency of LLMs at inference time and their ability to handle long sequences.
- –
- –
- –
- –
- –
- Measuring Cross-Lingual Transferability of Multilingual Transformers on Sentence Classification–
IGap, a cross-lingual transferability metric for multilingual Transformers on sentence classification tasks, is proposed, which takes training error into consideration, and can also estimate transferability without end-task data.
- S2F-NER: Exploring Sequence-to-Forest Generation for Complex Entity Recognition–
This paper proposes a novel Sequence-to-Forest generation paradigm, S2F-NER, which can directly extract entities in sentence via a Forest decoder that decode multiple entities in parallel rather than sequentially.
- CRAT-XML: Contrastive Representation Adversarial Training for Extremely Multi-Label Text Classification–
CRAT-XML is proposed, which uses contrast adversarial learning to optimize text representation and enhance the acquisition of dependency relations between text and labels, thus reducing the need for integration at the representation level and achieving relatively high accuracy under low-resource and low-time conditions.
- –
- –
- –
- –
- Prediction or Comparison: Toward Interpretable Qualitative Reasoning–
This work categorizes qualitative reasoning tasks into two types: prediction and comparison, and adopts neural network modules trained in an end-to-end manner to simulate the two reasoning processes.
- –
- Incorporating Domain Knowledge into Text Classification Diagnosis in Customer Service Dialogue Field–
A domain knowledge-based text classification diagnosis method, which innovatively transforms the question and answer task into the text classification task and uses an ERNIE-based structure to match consumer questions with multivariate groups of answers from the knowledge base, judged by similarity.
- Customer Satisfaction Research based on Customer Service Dialogue Corpus–
This work introduces an ERNIE-based customer satisfaction analysis model that automatically analyses the text of customer service dialogues and scores them from four perspectives (i.e., product, service, process and overall) without human involvement.
- –
- –
- Multi-level Chunk-based Constituent-to-Dependency Treebank Transformation for Tibetan Dependency Parsing–
This work proposes a novel method of multi-level chunk-based syntactic parsing to complete constituent-to-dependency treebank conversion for Tibetan under scarce conditions and embodies the regularity of strong knowledge-guided linguistic analysis methods.
- –
- –
- Deep Hashing with Attribute Guidance for Image Retrieval–
A novel supervised deep hashing method with image attribute guidance that is learnt through image visual features and guided by image attributes by maintaining pair wise similarities between images as well as the corresponding attribute descriptions.
- –
- –
- –
- –
- Picture News Collection: A Dataset for Automatic Picture News Thumbnail Selection–
The automatic picture news thumbnail selection task is introduced, an attention-based thumbnail selection model is proposed, and experimental results comparing with three image classification based baselines show that the proposed methods outperform the baselines.
- –
- –
- Topical Sentence Embedding for Query Focused Document Summarization.–
A novel sentence embedding framework to collaborate the current sentence representation, word-based content and topic assignment of the sentence to predict the next sentence representation is proposed and shows the model outperforms state-of-the-art methods.
- –
- –
- –
- –
- A Comparative Study of Answer-Contained Snippets and Traditional Snippets–
This paper proposes a simple answer-contained snippet method for community-based Question and Answer (cQA) search, and compares the method with the state-of-the-art traditional snippet algorithms.
- –
- –
- –
- Improving Conversational Spoken Language Machine Translation via Pronoun Recovery–
This work inserts the position of missing pronouns into the source side, then the word alignment method is used to filter the pronouns in order to pick up the pronouns which are really helpful for the machine translation.
- Multi-roles Affiliation Model for General User Profiling–
This work designs a novel, flexible general user profiling model which naturally models users’ friendships with user attributes and shows that this method simultaneously profile multiple attributes with better performance.
- Chinese Text Analysis Based on Minimal Recursion Semantics–
To represent Chinese texts using MRS, a lexicon is built with the rich semantic knowledge of HowNet and defined36 grammar rules and 47 types to show that MRS is also suitable for Chinese semantic representation.
- Topic-Based Chinese Message Polarity Classification System at SIGHAN8-Task2–
This paper describes the topic-based Chinese message polarity classification system submitted by LCYS_TEAM at SIGHAN8-Task2 and proposes a new weighting approach and a PMI-based random jumping probability selection method for sentimental features.
- –
- A Hybrid Method of Sentiment Key Sentence Identification Using Lexical Semantics and Syntactic Dependencies–
Experimental results on COAE 2014 dataset show that this approach notablely outperforms other baselines of sentiment key sentence identification and performs dependency parsing on the basis of a novel lexicon-based algorithm.
- Chinese Evaluation Phrase Extraction Based on Cascaded Model–
An efficient method for extracting Chinese evaluation phrase based on cascaded model and three kinds of rule-based methods, such as parenthesis/preposition/adverb phrase rule, are designed to extract evaluation phrase in complex structure is proposed.
- Learning to Display in Sponsored Search–
A novel dynamic algorithm is proposed, which is extended from the state-of-the-art multiple-choice secretary algorithm, and it is demonstrated that the algorithm outperforms the fundamental static algorithm significantly and can scale up with respect to very large datasets.
- A Unified Generative Model for Characterizing Microblogs’ Topics–
A unified generative model for characterizing microblogs' topics based on topic models is proposed and it is demonstrated that this approach consistently outperforms previous methods at a significant scale.
- Online Cross-Lingual PLSI for Evolutionary Theme Patterns Analysis–
This paper proposes an online cross-lingual topic model that can efficiently reduce the updating time complexity; and it is effective in solving the continuity limitation.
- –
- –
- Quantitative Omission Model of Candidate Unknown Words for Chinese Word Segmentation Based Repeat Extraction–
A quantitative omission model for candidate unknown words based on CWS is presented to evaluate the problem of omission of unknown words and a reliable conclusion of Chinese UWI via two strategies is reached, which has certain reference value for follow-up researches of UWI.
- Pretreatment for Speech Machine Translation–
A pretreatment approach which based on conditional random field model was introduced to delete the meaningless modal particles and dittographes, correct the recognition errors, and punctuated the ASR results before machine translation.
- –
- –
- –
- Translation & Transform Algorithm of Query Sentence in Cross-Language Information Retrieval–
This paper explores the application of the traditional monolingual IR technology to converting the translation of query sentence to the computation of the boost value of query keyword translations in the bilingual dictionary, so that the target language query sentence is reconstructed.
- Context Sensitive Automatic Chinese Word Segmentation and Lexical Preprocessing–
A context sensitive automatic Chinese word segmentation and lexical preprocessing for Chinese English machine translation system is proposed, taking advantage of large amount of syntax, semantic and common sense knowledge in the lexicon of MT system.
- Parsing with dynamic rule selection–
By describing the design and practice of the improved algorithm, the paper discusses in detail the designing method for a high speed and efficient parser.
- A parallel implementation model of HPARLOG–
A new parallel logic programming language—HPARLOG developed by us is described, and a new scheme for the AND-parallelism implementation in logic Programming language is proposed, which resolves the instantiation conflict on sharing-variables and decreases the dynamic complexity of the programs.
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-12. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.