Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks153 from public data
- GEAR: Graph-based Evidence Aggregating and Reasoning for Fact Verification249
A graph-based evidence aggregating and reasoning (GEAR) framework which enables information to transfer on a fully-connected evidence graph and then utilizes different aggregators to collect multi-evidence information is proposed.
- MAVEN: A Massive General Domain Event Detection Dataset232
A MAssive eVENt detection dataset is presented, which contains 4,480 Wikipedia documents, 117,200 event mention instances, and 207 event types and shows that existing ED methods cannot achieve promising results as on the small datasets, which suggests ED in real world remains a challenging task and requires further research efforts.
- SK-GCN: Modeling Syntax and Knowledge via Graph Convolutional Network for aspect-level sentiment classification210
A new Syntax- and Knowledge-based Graph Convolutional Network (SK-GCN) model is proposed for aspect-level sentiment classification, which leverages the syntactic dependency tree and commonsense knowledge via GCN to enhance the representation of the sentence toward the given aspect.
- RAP: Robustness-Aware Perturbations for Defending against Backdoor Attacks on NLP Models164
This work constructs a word-based robustness-aware perturbation to distinguish poisoned samples from clean samples to defend against the backdoor attacks on natural language processing (NLP) models.
- 152
- Deep Learning for Aspect-Level Sentiment Classification: Survey, Vision, and Challenges152
A taxonomy of deep learning-based aspect-level sentiment classification is designed and a comprehensive summary of the state-of-the-art methods is given, showing the tremendous progress that has already been made in ASC.
- Rethinking Stealthiness of Backdoor Attack against NLP Models151
A novel word-based backdoor attacking method based on negative data augmentation and modifying word embeddings is proposed, making an important step towards achieving stealthy backdoor attacking.
- CLEVE: Contrastive Pre-training for Event Extraction133
CLEVE is proposed, a contrastive pre-training framework for EE to better learn event knowledge from large unsupervised data and their semantic structures (e.g. AMR) obtained with automatic parsers.
- Improving Multi-turn Dialogue Modelling with Utterance ReWriter125
This paper proposes rewriting the human utterance as a pre-process to help multi-turn dialgoue modelling and introduces a Transformer-based utterance rewriting architecture using the pointer network, which achieves remarkably good performance on the utterance rewrite task.
- Cross-modal fine-grained alignment and fusion network for multimodal aspect-based sentiment analysis93
- 93
- SentiX: A Sentiment-Aware Pre-Trained Model for Cross-Domain Sentiment Analysis93
This paper pre-train a sentiment-aware language model (SentiX) via domain-invariant sentiment knowledge from large-scale review datasets, and utilize it for cross-domain sentiment analysis task without fine-tuning.
- A dependency syntactic knowledge augmented interactive architecture for end-to-end aspect-based sentiment analysis91
This paper proposes a novel dependency syntactic knowledge augmented interactive architecture with multi-task learning for end-to-end ABSA that significantly outperforms existing state-of-the-art methods.
- CM-Net: A Novel Collaborative Memory Network for Spoken Language Understanding90
A novel Collaborative Memory Network (CM-Net) based on the well-designed block, named CM-block, which achieves the state-of-the-art results on the ATIS and SNIPS in most of criteria, and significantly outperforms the baseline models on the CAIS.
- Recent Advances of Foundation Language Models-based Continual Learning: A Survey85
This article delves into a comprehensive review, summarization, and classification of the existing literature on CL-based approaches applied to foundation language models, such as pre-trained language models, large language models, and vision-language models.
- 79
- Conversations Are Not Flat: Modeling the Dynamic Information Flow across Dialogue Utterances68
The DialoFlow model is proposed, in which a dynamic flow mechanism to model the context flow is introduced, and three training objectives are designed to capture the information dynamics across dialogue utterances by addressing the semantic influence brought about by each utterance in large-scale pre-training.
- 62
- Position-aware hierarchical transfer model for aspect-level sentiment classification60
A position-aware hierarchical transfer (PAHT) model that models the position information from multiple levels and enhances the ASC performance by transferring hierarchical knowledge from the resource-rich sentence-level sentiment classification (SSC) dataset is proposed.
- A Novel Aspect-Guided Deep Transition Model for Aspect Based Sentiment Analysis49
A novel Aspect-Guided Deep Transition model, named AGDT, is proposed, which utilizes the given aspect to guide the sentence encoding from scratch with the specially-designed deep transition architecture, and an aspect-oriented objective is designed to enforce AGDT to reconstruct thegiven aspect with the generated sentence representation.
- Towards Fine-grained Text Sentiment Transfer47
This paper proposes a novel Seq2SentiSeq model, where the numeric sentiment intensity value is incorporated into the decoder via a Gaussian kernel layer to finely control the sentiment intensity of the output.
- 46
- An Iterative Multi-Knowledge Transfer Network for Aspect-Based Sentiment Analysis44
This work proposes a novel Iterative Multi-Knowledge Transfer Network (IMKTN) for end-to-end ABSA, which transfers the task-specific knowledge from any two of the three subtasks to another one at the token level by utilizing a well-designed routing algorithm.
- Boosting Conversational Question Answering with Fine-Grained Retrieval-Augmentation and Self-Check42
A conversation-level RAG (ConvRAG) approach, which incorporates fine-grained retrieval augmentation and self-check for conversational question answering (CQA) and a Chinese CQA dataset with new features, which facilitates further researches in RAG enhanced CQA.
- Contrastive Zero-Shot Learning for Cross-Domain Slot Filling with Adversarial Attack39
A Contrastive Zero-Shot Learning with Adversarial Attack (CZSL-Adv) method for the cross-domain slot filling that aims to map slot value contextual representations to the corresponding slot description representations and introduces an adversarial attack training strategy to improve model robustness.
- LLM-DA: Data Augmentation via Large Language Models for Few-Shot Named Entity Recognition38
This paper proposes $LLM-DA, a novel data augmentation technique based on LLMs for the few-shot NER task that leverages the distinctive characteristics of the NER task by augmenting the original data at both the contextual and entity levels.
- DelTA: An Online Document-Level Translation Agent Based on Multi-Level Memory38
Results indicate that DelTA significantly outperforms strong baselines in terms of translation consistency and quality across four open/closed-source LLMs and two representative document translation datasets, achieving an increase in consistency scores by up to 4.58 percentage points and in COMET scores by up to 3.16 points on average.
- GoG: Relation-aware Graph-over-Graph Network for Visual Dialog38
A novel relation-aware graph-over-graph network (GoG) for visual dialog that outperforms the strong baseline in both generative and discriminative settings by a significant margin.
- Boosting Large Language Models with Socratic Method for Conversational Mathematics Teaching37
This paper collects and releases a high-quality mathematical teaching dataset, named SocraticMATH, which provides Socratic-style conversations of problems with extra knowledge and proposes a knowledge-enhanced LLM as a strong baseline to generate reliable responses with review, guidance/heuristic, rectification, and summarization.
- DMRM: A Dual-Channel Multi-Hop Reasoning Model for Visual Dialog35
A novel and more powerful Dual-channel Multi-hop Reasoning Model for Visual Dialog, named DMRM, which synchronously captures information from the dialog history and the image to enrich the semantic representation of the question by exploiting dual-channel reasoning.
- 35
- Optimal answerer ranking for new questions in community question answering34
This paper utilizes tensor model and topic model simultaneously to extract latent semantic relations among asker, question and answerer and proposes a learning procedure to get optimal ranking of answerers for new questions by optimizing the multi-class AUC (Area Under the ROC Curve).
- 32
- Is position important? deep multi-task learning for aspect-based sentiment analysis31
Inspired by autoencoder structure, a position-aware attention and a deep bi-directional LSTM (DBi-LSTM) model for sentiment prediction and position prediction respectively to capture the position information better are designed.
- Mathematical Language Models: A Survey29
A comprehensive survey of mathematical LMs is conducted, systematically categorizing pivotal research endeavors from two distinct perspectives: tasks and methodologies.
- MSCTD: A Multimodal Sentiment Chat Translation Dataset29
A new task named Multimodal Chat Translation (MCT), aiming to generate more accurate translations with the help of the associated dialogue history and visual context, and provides a new benchmark on multimodal dialogue sentiment analysis with the constructed MSCTD.
- Multimodal Incremental Transformer with Visual Grounding for Visual Dialogue Generation27
A multi-turn dialogue history combined with visual scene step by step according to the order of the dialogue and then generates a contextually and visually coherent response, named MITVG, which achieves comparable performance on the VisDial datasets.
- Retrieval-Polished Response Generation for Chatbot26
A retrieval-polished model for response generation that polishes a draft response based on a retrieved prototype that outperforms retrieval-based and generation-based chatbots with respect to fluency, contextual relevance, and response diversity.
- Context Tracking Network: Graph-based Context Modeling for Implicit Discourse Relation Recognition25
A novel graph-based Context Tracking Network (CT-Net) to model the discourse context for implicit discourse relation recognition (IDRR), which shows that the CT-Net gains better performance than models that roughly model the context.
- Memory-based Adapters for Online 3D Scene Perception24
An adapter-based plug-and-play module for the backbone of 3D scene perception model, which constructs memory to cache and aggregate the extracted RGB-D features to empower offline models with temporal learning ability, and a queued memory mechanism to cache the supporting point cloud and image features.
- 21
- Improving Gradient-based Adversarial Training for Text Classification by Contrastive Learning and Auto-Encoder20
This work focuses on enhancing the model's ability to defend gradient-based adversarial attack during themodel's training process and proposes two novel adversarial training approaches: CARL and RAR, which outperform strong baselines on various text classification datasets and can effectively improve the robustness of the model.
- PG-RNN: using position-gated recurrent neural networks for aspect-based sentiment classification20
A Position-Gated Recurrent Neural Networks model that considered aspect word position information and uses kernel function to model position information instead of discrete distance values to integrate global and local information dynamically for aspect-based sentiment classification.
- CMM-Math: A Chinese Multimodal Math Dataset To Evaluate and Enhance the Mathematics Reasoning of Large Multimodal Models19
The extensive experiments indicate that the model effectively improves math reasoning performance by comparing it with the SOTA LMMs over three multimodal mathematical datasets, and proposes a Multimodal Mathematical LMM (Math-LMM) to handle the problems with mixed input of multiple images and text segments.
- 17
- 17
- SNNN: Promoting Word Sentiment and Negation in Neural Sentiment Classification17
LSTM performs better than CNN and RNN for neural sentiment classification; word sentiment and negation are a strong alliance with attention, while overfitting occurs when they are simultaneously applied at the embedding layer; and word sentiment/negation can be singly implemented for better performance as bothembedding layer and attention at the same time.
- Boosting e-commerce sales with live streaming: the power of barrages16
The results demonstrate an inverted U-shaped effect of barrage count and length on online product sales, and it is discovered that the inverted U-shaped curve reaches its turning point with increased barrage count and length.
- A Sentiment-Controllable Topic-to-Essay Generator with Topic Knowledge Graph15
Experimental results show that the novel Sentiment Controllable topic-to- essay generator with a Topic Knowledge Graph enhanced decoder, named SCTKG, can generate sentiment controllable essays and outperform the state-of-the-art approach in terms of topic relevance, fluency, and diversity on both automatic and human evaluation.
- Entity-level sentiment prediction in Danmaku video interaction14
A large-scale Chinese video comments dataset containing time-sync Danmaku comments and traditional video comments is proposed, targeting multiple entities and sentiments associated with each entity from popular video websites, and a framework of entity-level sentiment analysis with two de-biasing models: hard-masking de-bias model and soft-masked de-Bias model are proposed.
- Vocal Emotion of Humanoid Robots: A Study from Brain Mechanism14
This paper explored the brain mechanism of vocal emotion by studying previous researches and developed an experiment to observe the brain response by fMRI, to analyze vocal emotion of human beings.
- A New Approach to Answerer Recommendation in Community Question Answering Services14
Experimental results on two real-world CQA dataset show that the proposed method is able to recommend appropriate answerers for new questions and outperforms other state-of-the-art approaches.
- Enhancing Visual Dialog Questioner with Entity-based Strategy Learning and Augmented Guesser13
A Related entity enhanced Questioner (ReeQ) that generates questions under the guidance of related entities and learns entity-based questioning strategy from human dialogs and an Augmented Guesser (AugG) that is strong and is optimized for the VD setting especially are proposed.
- 13
- PT-BitNet: Scaling up the 1-Bit large language model with post-training quantization11
PT-BitNet is introduced, a novel post-training quantization method that extends the benefits of BitNet's ternary quantization to large-scale language models up to 70B parameters and achieves substantial reductions in model size and inference time, with minimal impact on task performance.
- 11
- 11
- Survey on Multi-Source Data Based Application and Exploitation Toward Smart Ship Navigation10
This paper comprehensively analyzes the key technologies of maritime navigation based on multi-source data and highlights the advantages and cutting-edge methods of multi-source data in typical application scenarios and analyze their performance and adaptation strategies in different usage contexts.
- 10
- 10
- 10
- UniTTA: Unified Benchmark and Versatile Framework Towards Realistic Test-Time Adaptation9
A versatile UniTTA framework is proposed, which includes a Balanced Domain Normalization layer and a COrrelated Feature Adaptation method--designed to mitigate distribution gaps in domain and class, respectively--designed to mitigate distribution gaps in domain and class.
- Modeling Multi-aspect Relationship with Joint Learning for Aspect-Level Sentiment Classification9
A joint learning (Joint) model for aspect-level sentiment classification is proposed, which models the relationships among the aspects of the sentence and predicts the sentiment polarities of all aspects simultaneously.
- A Soft Contrastive Learning-Based Prompt Model for Few-Shot Sentiment Analysis8
A sentiment-aware chain of thought prompt module is designed to guide the model to predict the sentiment from coarse grain to fine grain via a series of intermediate reasoning steps, and a soft contrastive learning algorithm is proposed to take the correlation of the labels into account.
- 8
- 8
- Enhancing Event-Level Sentiment Analysis with Structured Arguments7
This paper redefine the task as structured event-level SA and proposes an End-to-End Event-level Sentiment Analysis (E3SA) approach to solve this issue, and explicitly extract and model the event structure information for enhancing event- level SA.
- Attending via both Fine-tuning and Compressing7
This paper proposes a framework consisting of a learner and a compressor, which performs tuning and compressing iteratively to enhance the performance and interpretability of the attention mechanism.
- 6
- KACC: A Multi-task Benchmark for Knowledge Abstraction, Concretization and Completion6
This paper proposes large-scale datasets extracted from Wikidata, which provide more size-balanced concept graphs and abundant cross-view links and proposes a benchmark to test the ability of existing models on knowledge abstraction, concretization and completion (KACC).
- Modeling Explicit Concerning States for Reinforcement Learning in Visual Dialogue6
ECS is modeled from multimodal information and is represented explicitly and enables the Visual Dialogue agents to generate more visual coherent, less repetitive and more visual informative dialogues compared with previous methods, according to multiple automatic metrics, human study and qualitative analysis.
- Method for BBS topic tracking based on semantic similarity6
The paper utilized the characteristics of BBS and presented a topic tracking method for BBS data based on semantic similarity that effectively avoids the disadvantage of Vector Space Model (VSM).
- Enhancing the convolution-based knowledge graph embeddings by increasing dimension-wise interactions5
- 5
- A weakly supervised knowledge attentive network for aspect-level sentiment classification5
A Knowledge Attentive Network is presented to capture more aspect-related information by incorporating external commonsense knowledge into the attention mechanism and a weakly supervised learning method is proposed, which allows the KAN model to learn more knowledge from the pseudo-samples generated upon the rich-resource document-level sentiment classification datasets.
- 5
- 5
- 5
- Dissecting Long-Chain-of-Thought Reasoning Models: An Empirical Study4
A systematic analysis of the roles of positive and negative samples in RL reveals that positive samples mainly facilitate data fitting, whereas negative samples significantly enhance generalization and robustness, and identifies substantial data inefficiency in group relative policy optimization.
- 4
- 4
- 4
- 3
- 3
- Rethinking Overlooked Aspects in Vision-Language Models3
It is indicated that merely increasing the size of pre-training data does not guarantee improved performance and may, in fact, lead to its degradation, and a pipeline to pinpoint the most efficient instruction tuning dataset is established, implying that not all SFT data utilized in existing studies are necessary.
- 3
- A Knowledge-Enhanced Adversarial Model for Cross-lingual Structured Sentiment Analysis3
A Knowledge-Enhanced Adversarial Model (KEAM) with both implicit distributed and explicit structural knowledge to enhance the cross-lingual transfer is proposed, which outperforms all the unsupervised baselines in various metrics.
- Is “hot pizza” Positive or Negative? Mining Target-aware Sentiment Lexicons3
This paper proposes an effective “perturb-and-see” method to extract sentiment words modifying it from large-scale datasets and shows that a simple application of the lexicon is able to achieve highly competitive performances on the unsupervised opinion relation extraction task.
- 3
- Mobile Wikipedia: A Case Study of Information Service Design for Chinese Teenagers3
Results of this study have reference value for practitioners to involve target users into development process of information service and apply User Centered Design in mobile service design.
- 2
- TamEdit: Trajectory-Aware Meta-Learning for Specificity-Preserving Continual Knowledge Editing2
Experiments show TamEdit significantly outperforms baselines in continual editing, improving specificity by 14.81% with sub-second inference speed (0.55s per edit), while preserving general capabilities.
- 2
- Control-CLIP: Decoupling Category and Style Guidance in CLIP for Specific-Domain Generation2
Control-CLIP is proposed, a novel decoupled CLIP fine-tuning framework that enables the CLIP model to learn the meaning of category and style in a complement manner and can precisely guide the diffusion model to a specific domain.
- 2
- 2
- Visual Dialog for Spotting the Differences between Pairs of Similar Images2
This paper proposes a new visual dialog task named Dial-the-Diff, where two interlocutors accessing two similar images respectively try to spot the difference between the images through conversing in natural language, which raises new challenges to the dialog strategy and the ability of categorizing objects.
- Multi-Zone Unit for Recurrent Neural Networks2
A new Multi-zone Unit (MZU) for RNNs is introduced, to design a transition function that is capable of modeling multiple space composition and experimental results demonstrate the superiority of the MZU.
- Topic model with incremental vocabulary based on Belief Propagation2
In order to address the drawbacks of LDA, ivLDA with topic–word distribution stemming from the Dirichlet process that has infinite atoms instead ofDirichlet distribution is proposed and two methods are presented to manage the indices of the words, namely, iv LDA-Perp and ivL DA-PMI.
- ZZISTI at TREC2013 Temporal Summarization Track2
The team submitted runs for the first running of the TREC Temporal Summarization track, where Stanford CoreNLP was applied to extract the event attributes.
- 2
- An AI-Structural Approach to Edge Detection2
A new approach to edge detection based on the intuition of edge is presented, finding of a special data structure of tree to represent both gray level and structural information in picture conveniently.
- LLM-KT: Enhancing Large Language Models with Knowledge Tracing via Multi-Level Plug-and-Play Alignment1
LLM-KT is proposed, a novel framework that integrates the reasoning power of Large Language Models with the sequential modeling strengths of traditional KT methods via multi-level plug-and-play alignment and achieves state-of-the-art performance.
- ReFreeKV: Towards Threshold-Free KV Cache Compression1
A new objective is proposed that lifts the threshold constraints for robust KV compression, advocating for "threshold-free" methods that adaptively adjust budget allocation while preserving full-cache performance.
- 1
- 1
- Black-box Model Merging for Language-Model-as-a-Service with Massive Model Repositories1
A derivative-free optimization framework based on the evolutionary algorithm (Evo-Merging) that enables effective model merging using only inference-time API queries, and consists of sparsity-based denoising and sign-aware scaling, which dynamically computes optimal combination weights for the relevant models based on their performance.
- Large Language Models have Chain-of-Affect1
This work introduces the concept of the chain-of-affect (CoA), a temporally extended affective process through which LLMs develop state-like behavioral tendencies that shape generation, user experience, and collective dynamics.
- A Synthetic-to-Real Dehazing Method based on Domain Unification1
A synthetic-to-real dehazing method based on domain unification is proposed, which attempts to unify the relationship between the real and synthetic domain, thus to let the dehazing model more in line with the actual situation.
- 1
- 1
- DSPF: A Digital Signal Processing Based Framework for Information Retrieval1
This research proposes a new DSP-based IR Framework (DSPF) introducing the theories from the field of the DSP and presents two corresponding DSP -BM25 and DSPF-DLM models, which incorporate the term weighting schemes from two well-performed probabilistic IR models, the BM25 and the Dirichlet Language Model (DLM).
- 1
- 1
- 1
- Corrections to “Efficient Motion Vector Interpolation for Error Concealment of H.264/AVC” [Mar 11 75-80]1
In the above paper (ibid., vol. 57, no. 1, pp. 75-80, Mar. 2011), the financial support information was omitted from the first footnote, so it is provided here.
- 1
- 1
- –
- –
- –
- –
- –
- Investigating Cross-Modal Skill Injection: Scenarios, Methods, and Hyperparameters–
This study investigates cross-modal skill injection across three main aspects: scenarios, methods, and hyperparameters, and finds that cross-modal skill injection generally performs well in instruction-following and cross-lingual settings, yet struggles with mathematical reasoning.
- –
- FlowRAG: Synergizing Explicit Reasoning via Frequency-Aware Multi-Granularity Graph Flow–
This work constructs a quad-level heterogeneous graph over passages, summaries, sentences, and entities, where summary nodes serve as a coarse semantic hub and introduces a frequency-aware weighted flow module that routes relevance through entity--passage links weighted by within-passage term frequency.
- CoDoL: Conditional Domain Prompt Learning for Out-of-Distribution Generalization–
A novel Conditional Domain prompt Learning (CoDoL) method is proposed, which utilizes readily-available domain information to form prompts and contributes to improved vision-language embedding alignment, which is identified as one factor underlying the observed OOD generalization gains.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Modeling Inter-Aspect Dependencies with a Non-temporal Mechanism for Aspect-Based Sentiment Analysis–
- –
- RETRACTED: Analysis on the Innovation Hierarchy of Commercial Bank’s Financial based on the Big Data–
This paper analyzes the impact of big data on financial services by combining computer big data technology, and puts forward suggestions for small and micro-enterprise financial product innovation of commercial banks based on computerBig data in view of the current difficulties in financialProduct innovation ofCommercial banks.
- A New Digital Signal Processing Based Model With Multi-Aspect Term Frequency for Information Retrieval–
This research proposes a new D SP-based IR model, denoted as DSP-MATF, which incorporates a very effective term weighing method from the well-performed Vector Space Model named Multi-Aspect Term Frequency (MATF).
- –
- –
- –
- –
- –
- –
- –
- Rating of web pages based on image content analysis–
A novel framework to rate webpage using image content analysis is proposed, which is in compliance with the standard utilized by most web browsers like Internet Explorer, which include Normal, Revealing Attire, Exposed Breasts and Bare Buttocks.
- –
- –
- –
- –
- –
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 2026-10-15. 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.