Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Ting LiuSuggested from co-authorship
- Guohong FuSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works167 from public data
- Unified Named Entity Recognition as Word-Word Relation Classification392
This work develops a neural framework, in which the unified NER is modeled as a 2D grid of word pairs, and proposes multi-granularity 2D convolutions for better refining the grid representations.
- Gated Neural Networks for Targeted Sentiment Analysis311
A sentence-level neural model is proposed to address the limitation of pooling functions, which do not explicitly model tweet-level semantics and gives significantly higher accuracies compared to the current best method for targeted sentiment analysis.
- A neural joint model for entity and relation extraction from biomedical text251
It is demonstrated that the model based on neural networks is effective for biomedical entity and relation extraction and parameter sharing is an alternative method for neural models to jointly process this task.
- Neural Networks for Open Domain Targeted Sentiment247
This work empirically studies the effect of word embeddings and automatic feature combinations on the open domain targeted sentiment task by extending a CRF baseline using neural networks, and proposes a novel integration of neural and discrete features which combines their relative advantages, leading to significantly higher results compared to both baselines.
- Tweet Sarcasm Detection Using Deep Neural Network217
This work investigates the use of neural network for tweet sarcasm detection, and compares the effects of the continuous automatic features with discrete manual features.
- GME: Improving Universal Multimodal Retrieval by Multimodal LLMs203
The General Multimodal Embedder (GME) is developed, an MLLM-based dense retriever designed for UMR, and a comprehensive UMR Benchmark (UMRB) is constructed to evaluate the effectiveness of this approach.
- End-to-End Neural Relation Extraction with Global Optimization188
This work builds a globally optimized neural model for end-to-end relation extraction, proposing novel LSTM features in order to better learn context representations, and presents a novel method to integrate syntactic information to facilitate global learning.
- Context-Sensitive Twitter Sentiment Classification Using Neural Network151
A context-based neural network model for Twitter sentiment analysis is proposed, incorporating contextualized features from relevant Tweets into the model in the form of word embedding vectors.
- 144
- Transition-Based Neural Word Segmentation126
Experimental results demonstrate that word features lead to comparable performance to the best systems in the literature, and a further combination of discrete and neural features obtains top accuracies on several benchmarks.
- A Span-Based Model for Joint Overlapped and Discontinuous Named Entity Recognition123
Experimental results on multiple benchmark datasets show that the proposed span-based model can be regarded as a relation extraction paradigm essentially and is highly competitive for overlapped and discontinuous NER.
- Cross-Lingual Semantic Role Labeling with High-Quality Translated Training Corpus116
This paper proposes a novel alternative based on corpus translation, constructing high-quality training datasets for the target languages from the source gold-standard SRL annotations, and Experimental results on Universal Proposition Bank show that the translation-based method is highly effective, and the automatic pseudo datasets can improve the target-language SRL performances significantly.
- LasUIE: Unifying Information Extraction with Latent Adaptive Structure-aware Generative Language Model114
In this work, a novel structure-aware GLM is proposed, fully unleashing the power of syntactic knowledge for UIE, and rich task-adaptive structural bias is learned that greatly resolves the UIE crux, the long-range dependence issue and boundary identifying.
- Improving Twitter Sentiment Classification Using Topic-Enriched Multi-Prototype Word Embeddings113
Two neural networks are developed which learning topic-enriched multi-prototype word embeddings (TMWE) that better capture tweet context by incorporating topic information and a single distributed representation for each word.
- HiTrans: A Transformer-Based Context- and Speaker-Sensitive Model for Emotion Detection in Conversations109
A transformer-based context- and speaker-sensitive model for EDC, namely HiTrans, which consists of two hierarchical transformers, which outperforms previous state-of-the-art models.
- 98
- Enhancing Video-Language Representations With Structural Spatio-Temporal Alignment87
Finnsta is designed as a plug&play system, which can be integrated into existing well-trained VLMs for further representation augmentation, without training from scratch or relying on SG annotations in downstream applications.
- On the Robustness of Aspect-based Sentiment Analysis: Rethinking Model, Data, and Training85
This study strengthens the current best-robust syntax-aware models by further incorporating the rich external syntactic dependencies and the labels with aspect simultaneously with a universal-syntax graph convolutional network and proposes to automatically induce high-quality synthetic training data with various types, allowing models to learn sufficient inductive bias for better robustness.
- 80
- Transition-based Neural RST Parsing with Implicit Syntax Features77
This paper proposes an implicit syntax feature extraction approach, using hidden-layer vectors extracted from a neural syntax parser, using a simple transition-based model as the baseline, further enhancing it with dynamic oracle.
- Character-Level Chinese Dependency Parsing76
This paper presents novel adaptations of two major shift-reduce dependency parsing algorithms to character-level parsing, and demonstrates improved performances over word-based parsing methods on the Chinese Treebank.
- Combining Discrete and Neural Features for Sequence Labeling74
The results on standard benchmarks show that state-of-the-art neural models can give accuracies comparable to the best discrete models in the literature for most tasks and combing discrete and neural features unanimously yield better results.
- 74
- Have You been Properly Notified? Automatic Compliance Analysis of Privacy Policy Text with GDPR Article 1373
This work targets to solve the problem of compliance analysis between GDPR (Article 13) and privacy policies and manually curates a corpus of 36,610 labeled sentences from 304 privacy policies, and conducts a rule-based analysis and a user study to evaluate the usability of the approach.
- 68
- SEE: Syntax-Aware Entity Embedding for Neural Relation Extraction66
This paper proposes to learn syntax-aware entity embedding for neural relation extraction, and encodes the context of entities on a dependency tree as sentence-level entities embedding based on tree-GRU.
- Syntax-Enhanced Neural Machine Translation with Syntax-Aware Word Representations62
This work proposes a novel method to integrate source-side syntax implicitly for NMT by using the intermediate hidden representations of a well-trained end-to-end dependency parser, referred to as syntax-aware word representations (SAWRs).
- Joint models for extracting adverse drug events from biomedical text61
Experimental results on a standard ADE corpus show that the discrete joint model outperforms a state-of-the-art baseline pipeline significantly and when discrete features are replaced by neural features, the recall is further improved.
- 58
- 56
- 54
- Transition-Based Disfluency Detection using LSTMs51
This model incrementally constructs and labels the disfluency chunk of input sentences using a new transition system without syntax information, which can capture non-local chunk-level features and is free for noise in syntax.
- Recognizing irregular entities in biomedical text via deep neural networks49
A neural network model called Bi-LSTM-CRF that consists of bidirectional (Bi) long short-term memories (LSTMs) and conditional random fields (CRFs) to identify regular entities and the components of irregular entities and it is shown that the method of automatically combining the components is as effective as the methods of manually designing rules.
- KaLM-Embedding-V2: Superior Training Techniques and Data Inspire A Versatile Embedding Model48
The proposed KaLM-Embedding-V2 series achieves state-of-the-art performance on the Massive Text Embedding Benchmark, outperforming models of comparable size and rivaling models 3--26x larger, setting a new standard for versatile and compact embedding models under 1B parameters.
- Syntax-Aware Neural Semantic Role Labeling47
Experiments on the benchmark CoNLL-2005 dataset show that syntax-aware SRL approaches can effectively improve performance over a strong baseline with external word representations from ELMo, and investigate several previous approaches for encoding syntactic trees to make a thorough study on whether extrantax-aware representations are beneficial for neural SRL models.
- AISHELL-NER: Named Entity Recognition from Chinese Speech46
The results demonstrate that the performance of NER from Chinese speech could be improved by combining entity-aware ASR and pretrained NER tagger, which can be easily applied to the modern SLU pipeline.
- 45
- End-to-end Semantic Role Labeling with Neural Transition-based Model45
This paper presents the first work of transition-based neural models for end-to-end SRL, and suggests high-order compositions to extract non-local features, which can enhance the proposed transition model further.
- Cross-Lingual Dependency Parsing Using Code-Mixed TreeBank44
Experiments on University Dependency Treebanks show that code-mixed treebanks are more effective than translated treebanks, giving highly competitive performances among cross-lingual parsing methods.
- 43
- Mastering the Explicit Opinion-Role Interaction: Syntax-Aided Neural Transition System for Unified Opinion Role Labeling43
This work investigates a novel solution by revisiting the transition architecture, and augmenting it with a pointer network (PointNet), and proposes a unified dependency-opinion graph (UDOG), co-modeling the syntactic dependency structure and the partial opinion-role structure.
- Retrieval-style In-context Learning for Few-shot Hierarchical Text Classification42
This work introduces the first ICL-based framework with LLM for few-shot HTC, and trains a pretrained language model with masked language modeling (MLM), layer-wise classification (CLS, specifically for HTC), and a novel divergent contrastive learning (DCL, mainly for adjacent semantically similar labels) objective.
- A Two-Stage Adaptation of Large Language Models for Text Ranking41
This work proposes a two-stage progressive paradigm to better adapt LLMs to text ranking, and performs SFT, and proposes an improved optimization strategy building upon RankLLaMA.
- A survey of syntactic-semantic parsing based on constituent and dependency structures40
This article briefly reviews the representative models of constituent parsing and dependency parsing, and also dependency graph parsing with rich semantics, and reviews the closely-related topics such as cross-domain, cross-lingual and joint parsing models, parser application as well as corpus development of parsing.
- Deep Learning in Sentiment Analysis39
An overview of the successful deep learning approaches sentiment analysis tasks at different levels is given, including sentiment classification, opinion extraction, fine-grained sentiment analysis, etc.
- Grounded Entity-Landmark Adaptive Pre-training for Vision-and-Language Navigation36
A novel Grounded Entity-Landmark Adaptive (GELA) pre-training paradigm for VLN tasks achieves state-of-the-art results on both tasks, demonstrating its effectiveness and generalizability.
- Cross-Lingual Semantic Role Labeling With Model Transfer35
This article proposes an end-to-end SRL model that incorporates a variety of universal features and transfer methods, and finds that gold syntax features are much more crucial for cross-lingual SRL, compared with the automatically-generated ones.
- Adversarial Learning for Chinese NER From Crowd Annotations35
This paper proposes an approach to performing crowd annotation learning for Chinese Named Entity Recognition (NER) to make full use of the noisy sequence labels from multiple annotators, Inspired by adversarial learning.
- Event-Driven Headline Generation32
This work proposes an event-driven model for headline generation that can be viewed as a novel combination of extractive and abstractive headline generation, combining the advantages of both methods using event structures.
- Constructing Holistic Spatio-Temporal Scene Graph for Video Semantic Role Labeling31
A novel holistic spatio-temporal scene graph (namely HostSG) representation based on the existing dynamic scene graph structures, which well model both the fine-grained spatial semantics and temporal dynamics of videos for VidSRL is explored.
- Joint POS Tagging and Dependence Parsing With Transition-Based Neural Networks31
Three neural network based classifiers are designed to resolve shift/reduce, tagging, and labeling conflicts in part-of-speech tagging and dependency parsing using transition-based neural networks.
- A Transition-based Model for Joint Segmentation, POS-tagging and Normalization28
Experimental results show that the proposed transition-based model can help improve the performance of word segmentation on microblogs, giving an error reduction in segmentation accuracy of 12.02%, compared to the traditional approach.
- RST Discourse Parsing with Second-Stage EDU-Level Pre-training27
This work takes a state-of-the-art transition-based neural parser as baseline, and adopts it with a light bi-gram EDU modification to effectively explore the EDU-level pre-trained EDU representation.
- Uni-MoE-2.0-Omni: Scaling Language-Centric Omnimodal Large Model with Advanced MoE, Training and Data26
The Uni-MoE 2.0-Omni model substantially advances Lychee's Uni-MoE series in language-centric multimodal understanding, reasoning, and generating and achieves SOTA or highly competitive performance against leading OLMs, surpassing Qwen2.5-Omni on over 50 of 76 benchmarks.
- LLM-Assisted Data Augmentation for Chinese Dialogue-Level Dependency Parsing26
This study focuses on Chinese dialogue-level dependency parsing, presenting three simple and effective strategies with LLM to augment the original training instances, namely word-level, syntax-level, and discourse-level augmentations, respectively.
- Dependency-based syntax-aware word representations26
The proposed method can outperform the Tree-RNN and Tree-Linearization approaches in most settings, and meanwhile are highly efficient in syntax integration, and would be easily extendable to encoding other structural attributes of language.
- End-to-end neural opinion extraction with a transition-based model24
A novel end-to-end neural model alternatively for joint opinion extraction is proposed, by using a transition-based framework that exploits multi-layer bi-directional long short term memory networks to encode the input sentences, and then decode incrementally based on partial output results dominated by a transition system.
- Making Decision like Human: Joint Aspect Category Sentiment Analysis and Rating Prediction with Fine-to-Coarse Reasoning23
Experimental results demonstrate that the proposed framework outperforms state-of-the-art baselines by large margins and in-depth analyses prove that the method is effective on addressing both the unbalanced data distribution and the long-text issue.
- Conversational Semantic Role Labeling with Predicate-Oriented Latent Graph23
This work proposes to automatically induce a predicate-oriented latent graph (POLar) with apredicate-centered gaussian mechanism, by which the nearer and informative words to the predicate will be allocated with more attention.
- Domain-Specific NER via Retrieving Correlated Samples22
This paper draws correlated samples by the sparse BM25 retriever from large-scale in-domain unlabeled data and performs a training-free entity type calibrating by majority voting to simulate the human reasoning process and suggests enhancing NER models with correlated samples.
- 22
- Crowdsourcing Learning as Domain Adaptation: A Case Study on Named Entity Recognition22
This work investigates both unsupervised and supervised crowdsourcing learning, and suggests an annotator-aware representation learning model that inspired by the domain adaptation methods which attempt to capture effective domain-aware features.
- 22
- 21
- 20
- 19
- Study of compositions of musks in different types secreted by forest musk deer (Moschus berezovskii)19
The components were subjected to NMDS analysis to understand the differences in components among different types of musk, with the results suggesting that white musk was different from normal and black musks.
- Enhancing Opinion Role Labeling with Semantic-Aware Word Representations from Semantic Role Labeling18
A simple and novel method to enhance ORL by utilizing SRL is proposed, presenting semantic-aware word representations which are learned from SRL, which are fed into a baseline neural ORL model as basic inputs.
- APPCorp: a corpus for Android privacy policy document structure analysis17
This work creates a manually labelled corpus containing 231 privacy policies (of more than 566,000 words and 7,748 annotated paragraphs) and benchmarks the data corpus with 3 document classification models and achieves more than 82% on F1-score.
- 16
- LibN3L:A Lightweight Package for Neural NLP15
A light-weight machine learning tool for NLP research that provides several basic layers which mainly aims for single-layer linear and non-linear transformations, facilitating implementation of linear models as well as neural models.
- 14
- A Graph-Based Neural Model for End-to-End Frame Semantic Parsing14
This work exploits a graph-based method, regarding frame semantic parsing as a graph construction problem, and proposes an end-to-end neural model to tackle the task jointly, resulting in better performance than pipeline models.
- Communication of Uncertainty about Preliminary Evidence and the Spread of Its Inferred Misinformation during the COVID-19 Pandemic—A Weibo Case Study13
Investigating how the communication of uncertainty about preliminary evidence affects the spread of its inferred misinformation in a Weibo case study showed that attitude ambiguity toward the preliminary evidence and the stage when the evidence was first released with uncertainty were associated with higher numbers of likes and retweets of misinformation posts.
- Identifying personality traits and their potential application to the management of captive forest musk deer (Moschus berezovskii)13
The unfamiliar human being was the strongest stimulus to test boldness, followed by leopard feces, a beach ball and a leopard dummy, and nearly three quarter of tested individuals were identified as shy, while only one quarter was classified as bold.
- A False Trail to Follow: Differential Effects of the Facial Feedback Signals From the Upper and Lower Face on the Recognition of Micro-Expressions13
A novel hypothesis that facial feedback from upper and lower facial regions has differential effects on micro-expression recognition was derived and supported across three studies and highlighted the role of facial feedback in judging the subtle movements of micro-expressions.
- Stance Detection in Chinese MicroBlogs with Neural Networks13
A stance detection system for NLPCC-ICCPOL 2016 share task 4 can determinate whether the author of Weibo text is in favor of the given target, against thegiven target, or neither.
- In-Context Learning for Few-Shot Nested Named Entity Recognition12
This work introduces an effective and innovative ICL framework for the setting of few-shot nested NER, and improves the ICL prompt by devising a novel example demonstration selection mechanism, EnDe retriever.
- 12
- Combining Discrete and Continuous Features for Deterministic Transition-based Dependency Parsing12
By analyzing the results of various parsers on web-domain parsing, it is shown that the integrated model is a better way to combine traditional and embedding features compared with previous methods.
- 11
- BrainECHO: Semantic Brain Signal Decoding through Vector-Quantized Spectrogram Reconstruction for Whisper-Enhanced Text Generation10
BrainECHO demonstrates robustness across sentence, session, and subject-independent conditions, passing Gaussian noise tests and showcasing its potential for enhancing language-based brain-computer interfaces.
- DuAGNet: an unrestricted multimodal speech recognition framework using dual adaptive gating fusion10
To preserve the unique feature expression of each speech modality and improve the global perception of the coupling correlations among them, a Dual Adaptive Gating fusion framework is proposed (dubbed DuAGNet), utilizing modality-specific and feature-specific adaptive gating networks.
- Beyond Chunking: Discourse-Aware Hierarchical Retrieval for Long Document Question Answering9
A discourse-aware hierarchical framework that leverages rhetorical structure theory (RST) for long document question answering and converts discourse trees into sentence-level representations and employs LLM-enhanced node representations to bridge structural and semantic information.
- 9
- Finetuning Language Models for Multimodal Question Answering9
A T5-based multi-modal generative network is implemented that overcomes the limitations of traditional labeling space and provides more freedom in responses in the VTQA challenge.
- 9
- 9
- 9
- Segmenting Chinese Microtext: Joint Informal-Word Detection and Segmentation with Neural Networks9
This work proposes a joint segmentation model that integrates the detection of informal words simultaneously and generates training corpus for the joint model by using existing corpus automatically.
- 9
- Supervised Fine-Tuning or Contrastive Learning? Towards Better Multimodal LLM Reranking8
This work conducts a comprehensive comparison and analysis between CL and SFT for reranking, taking the universal multimodal retrieval (UMR) as the experimental playground and finds that SFT provides a substantially stronger weighting scheme than CL, whereas the preferred scoring direction shows no clear winner.
- SpeechEE: A Novel Benchmark for Speech Event Extraction8
This paper pioneers a SpeechEE, defined as detecting the event predicates and arguments from a given audio speech, and tailor an E2E SpeechEE system based on the encoder-decoder architecture, where a novel Shrinking Unit module and a retrieval-aided decoding mechanism are devised.
- 8
- A Fine-Grained Domain Adaption Model for Joint Word Segmentation and POS Tagging8
This work starts from joint word segmentation and POS tagging, presenting a fine-grained domain adaption method to model the gaps accurately, and proposes a novel domain-mixed representation learning model accordingly to encode the multiple subdomains effectively.
- Chinese Opinion Role Labeling with Corpus Translation: A Pivot Study8
This paper presents the first work of Chinese ORL, constructing a Chinese dataset by manually translating and projecting annotations from a standard English MPQA dataset, and investigates the effectiveness of cross-lingual transfer methods, including model transfer and corpus translation.
- DRTS Parsing with Structure-Aware Encoding and Decoding8
This work proposes a structural-aware model at both the encoder and decoder phase to integrate the structural information, where graph attention network (GAT) is exploited for effectively modeling.
- Semantic Role Labeling from Chinese Speech via End-to-End Learning7
This work presents the first end-to-end approach for SRL from speech, integrating ASR and SRL in a joint-learning framework, focusing on the Chinese language, and demonstrates the superiority of the end-to-end model over traditional pipelines, with significantly improved performance.
- Towards Demonstration-Aware Large Language Models for Machine Translation7
A novel approach that integrates demonstration-aware training and inference strategies within the framework of tuning-based LTMs, hereby referred to as demonstration-aware LTMs, which mitigates the negative impacts traditionally associated with demonstrations and secures substantial improvements in translation accuracy, particularly in domain-specific and document-level translation tasks.
- 7
- 7
- 6
- LLM-Driven Multimodal Opinion Expression Identification6
This study introduces a novel multimodal OEI (MOEI) task, integrating text and speech to mirror real-world scenarios, and proposes an LLM-driven method STOEI, which combines speech and text modal to identify opinion expressions.
- Nutritional status of early oral feeding for gastric cancer patients after laparoscopic total gastrectomy: A retrospective cohort study6
Among GC patients after LTG, EOF and traditional tube feeding had a similar risk of postoperative nutritional status, however, EOF was associated with a lower risk of abdominal distension.
- 6
- 6
- APGN: Adversarial and Parameter Generation Networks for Multi-Source Cross-Domain Dependency Parsing6
This work proposes a novel model for multi-source cross-domain dependency parsing that consists of a parameter generation network for distinguishing domain-specific features, and an adversarial network for learning domain-invariant representations.
- 5
- 5
- 5
- End to End Chinese Lexical Fusion Recognition with Sememe Knowledge5
This paper introduces Chinese lexical fusion recognition in detail and proposes an end-to-end model, handling mentions as well as coreference relationship jointly, which exploits the state-of-the-art contextualized BERT representations as an encoder and is further enhanced with the sememe knowledge from HowNet by graph attention networks.
- Domain Information Enhanced Dependency Parser5
This paper adopts the Bi-Affine parser as a baseline, and applies an ensemble-style self-training and adversarial learning to explore domain-specific information and domain-independent information for cross-domain dependency parsing.
- A Neural Joint Model for Extracting Bacteria and Their Locations5
A state-of-the-art neural joint model for relation extraction in the Automatic Content Extraction (ACE) task is adapted and a novel relation is suggested in the second step to detect the errors in the first step, thus this relation can correct some errors inThe first step.
- Dependency Tree Representations of Predicate-Argument Structures5
A novel annotation framework for representing predicate-argument structures that uses dependency trees to encode the syntactic and semantic roles of a sentence simultaneously is presented, which eliminates the structural gap between syntax and shallow semantics, making them compatible.
- Jointly or Separately: Which is Better for Parsing Heterogeneous Dependencies?5
This paper proposes a joint model which can make full use of the correlations between heterogeneous dependencies, and can answer the following question: parsing heterogeneous Dependency Parsing jointly or separately, which is better?
- When Words Smile: Generating Diverse Emotional Facial Expressions from Text4
This work introduces an end-to-end text-to-expression model that explicitly focuses on emotional dynamics and learns expressive facial variations in a continuous latent space and generates expressions that are diverse, fluid, and emotionally coherent.
- Chinese Sequence Labeling with Semi-Supervised Boundary-Aware Language Model Pre-training4
To assess PLMs’ ability to encode boundaries, a novel “Boundary Information Metric” is introduced that is both simple and effective and offers a convenient and accurate means of evaluating PLMs’ boundary awareness.
- 4
- Enhancing RDF Verbalization with Descriptive and Relational Knowledge4
Experimental results on the WebNLG and SemEval-2010 datasets show that the two types of knowledge can both enhance the model performance, and their combination is able to obtain further improvements in most cases, providing new state-of-the-art results.
- 4
- On the Role of Pre-trained Language Models in Word Ordering: A Case Study with BART4
To explain why BART helps word ordering, analysis is extended with probing and empirically identified that syntactic dependency knowledge in BART is a reliable explanation.
- Stacking Heterogeneous Joint Models of Chinese POS Tagging and Dependency Parsing4
A constituent-based joint model which derives the POS tag sequence and dependency tree from the output of PCFG parsers, and then integrate it into the guided transition- based joint model achieves the best performance.
- 3
- A Semantics Oriented Grammar for Chinese Treebanking3
This paper introduces a semantics oriented grammar for Chinese, designed to provide basic supports for tasks such as automatic semantic parsing and sentence generation, and is the first Chinese grammar representation capable of direct transformation into logical forms.
- Syntactic Processing Using Global Discriminative Learning and Beam-Search Decoding3
This tutorial discusses a framework for incremental left-to-right structured predication, which makes use of global discriminative learning and beam-search decoding, and gives an introduction to the algorithms and efficient implementations.
- A Comparison Study of Sequence Labeling Methods for Chinese Word Segmentation,POS Tagging Models3
Three different Chinese word segmentation and POS tagging models are compared and the joint model achieves the fastest speed while the stacked learning model achieves the highest accuracy.
- An End-to-End Model for Photo-Sharing Multi-Modal Dialogue Generation2
This work proposes the first end-to-end model for photo-sharing multi-modal dialogue generation, which integrates an image perceptron and an image generator with a large language model and employs the vision encoder to perceive visual images in the input end.
- 2
- 2
- 2
- Synergistic Dual Spatial-aware Generation of Image-to-Text and Text-to-Image2
This work considers modeling the SI2T and ST2I together under a dual learning framework, and proposes the Spatial Dual Discrete Diffusion (SD) framework, which utilizes the intermediate features of the 3D processes to guide the hard X$\to$3D processes, such that the overall ST2I and SI2T will benefit each other.
- Multi-Task Learning Network for Emotion Recognition in Conversation2
This paper exploits speaker identification (SI) as an auxiliary task to enhance the utterance representation in conversations, and can learn better speaker-aware contextual representations from the additional SI corpus.
- Cross-Lingual Dependency Parsing via Self-Training2
Results show that self-training can boost the dependency parsing performances on the target languages and the POS tagger assistant instance selection can achieve further improvements consistently.
- Chinese Sentiment Analysis Exploiting Heterogeneous Segmentations2
This paper studies the influences of these heterogeneous segmentations for Chinese sentiment analysis, and then integrates these segmentations, based on both discrete and neural models.
- Multiple TreeBanks Integration for Chinese Phrase Structure Grammar Parsing Using Bagging2
Bagging is proposed to enhance the baseline performance of PSG parsing in CIPS-Bakeoff2012 Task3.0 and outperforms the best closed system by nearly 4% which uses a single model.
- Micro blogs Oriented Word Segmentation System2
A Chinese word segmentation system submitted to the first task on CLP 2012 back-offs is presented using a conditional random field sequence model and derives a micro blog specified lexicon from auto-analyzed data and use lexicon related features to assist the model.
- Dependency Scoring Learning and Corpus Boosting for Translation-Based Cross-Lingual Dependency Parsing1
An iterative algorithm is designed to boost the target treebank quality, which trains with high-quality dependencies and relabels the low-quality dependencies and achieves better results than the originally translated treebanks and shows highly competitive performances with prior methods.
- 1
- BeMERC: Behavior-Aware MLLM-based Framework for Multimodal Emotion Recognition in Conversation1
A novel behavior-aware MLLM-based framework (BeMERC) is proposed to incorporate speaker's behaviors, including subtle facial micro-expression, body language and posture, into a vanilla MLLM-based MERC model, thereby facilitating the modeling of emotional dynamics during a conversation.
- 1
- Cross-domain Constituency Parsing by Leveraging Heterogeneous Data1
Experimental results show that the knowledge transfer model outperforms various baseline models, including conventional chart-based and transition-based parsers and the current large-scale language model for zero-shot and few-shot settings.
- 1
- Fine-Grained Domain Adaptation for Chinese Syntactic Processing1
Fine-grained domain adaptation for Chinese syntactic processing is proposed, aiming to model the gaps between the source and the target domains accurately and progressively, and an intuitive theoretical illustration based on the theory of Kumar et al. (2020) is offered.
- Chinese Event Discourse Deixis Resolution: Design of the Dataset and Model1
A deep neural network model is proposed for EDDR, which formulates the task into two subtasks, namely event anaphor recognition and event antecedent recognition, and outperforms all the baselines and achieves about 53%, 44%, 53%, and 63% F1s using standard anaphora resolution metrics.
- 1
- 1
- 1
- Discrete and Neural Models for Chinese POS Tagging: Comparison and Combination1
This paper compares the two kinds of models empirically, and incorporates word clusters for both models as well as their combination, since it has been generally accepted that word clusters can encode similar information as pre-trained word embeddings.
- Incremental Structured Prediction Using a Global Learning and Beam-Search Framework1
This tutorial discusses a framework for incremental left-to-right structured predication, which makes use of global discriminative learning and beam-search decoding, and gives an introduction to the algorithms and efficient implementations.
- 1
- A survey on fake review detection: From pre-trained language models to large language models–
This survey reviews fake review detection from an information fusion perspective, covering 211 studies published from 2018 to early 2026, and identifies open problems in adversarial generation, cross-domain transfer, uncertainty-aware fusion, missing-source robustness, interpretability, and trustworthy evaluation for AI-generated deceptive content.
- –
- –
- –
- –
- –
- –
- –
- The ACM Multimedia 2024 Viual Spatial Description Grand Challenge–
The VSD challenge provides two benchmark datasets for three subtasks, i.e., visual spatial relationship classification, single spatial description generation, and open-ended spatial description generation.
- –
- Cross-domain Chinese Sentence Pattern Parsing–
Experiments conducted on textbook and news domains demonstrate the effectiveness of the proposed method, outperforming rule-based baselines by 1.68 points on F1 metrics.
- –
- –
- –
- –
- Extracting Chinese Explanatory Expressions with Discrete and Neural CRFs–
This paper studies the extraction of explanatory expressions, by modeling the problem based on conditional random field (CRF), and compares the effectiveness of both discrete and neural features, and further integrate them.
- –
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.