Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks142 from public data
- 1,508
- Target-dependent Twitter Sentiment Classification1,117
This paper proposes to improve target-dependent Twitter sentiment classification by incorporating target- dependent features; and taking related tweets into consideration; and according to the experimental results, this approach greatly improves the performance of target- dependence sentiment classification.
- 861
- Improved Neural Relation Detection for Knowledge Base Question Answering322
This paper proposes a hierarchical recurrent neural network enhanced by residual learning which detects KB relations given an input question and proposes a simple KBQA system that integrates entity linking and the proposed relation detector to make the two components enhance each other.
- Improving Lexical Embeddings with Semantic Knowledge316
This work proposes a new learning objective that incorporates both a neural language model objective (Mikolov et al., 2013) and prior knowledge from semantic resources to learn improved lexical semantic embeddings.
- One-Shot Relational Learning for Knowledge Graphs293
This work proposes a one-shot relational learning framework, which utilizes the knowledge distilled by embedding models and learns a matching metric by considering both the learned embeddings and one-hop graph structures.
- R3: Reinforced Ranker-Reader for Open-Domain Question Answering283
This paper proposes a new pipeline for open-domain QA with a Ranker component, which learns to rank retrieved passages in terms of likelihood of extracting the ground-truth answer to a given question, and proposes a novel method that jointly trains the Ranker along with an answer-extraction Reader model, based on reinforcement learning
- Cross-lingual Knowledge Graph Alignment via Graph Matching Neural Network268
The topic entity graph is introduced, a local sub-graph of an entity, to represent entities with their contextual information in KG, and a graph-attention based solution is proposed that outperforms previous state-of-the-art methods by a large margin.
- Differential Treatment for Stuff and Things: A Simple Unsupervised Domain Adaptation Method for Semantic Segmentation256
The reason why the current adversarial loss is often unstable in minimizing the distribution discrepancy is revealed and it is shown that the method can help ease this issue by minimizing the most similar stuff and instance features between the source and the target domains.
- Diverse Few-Shot Text Classification with Multiple Metrics228
This work proposes an adaptive metric learning approach that automatically determines the best weighted combination from a set of metrics obtained from meta-training tasks for a newly seen few-shot task.
- StoryBuddy: A Human-AI Collaborative Chatbot for Parent-Child Interactive Storytelling with Flexible Parental Involvement219
StoryBuddy is an AI-enabled system for parents to create interactive storytelling experiences and highlighted the need for accommodating dynamic user needs between the desire for parent involvement and parent-child bonding and the goal of minimizing parent intervention when busy.
- Improved Relation Extraction with Feature-Rich Compositional Embedding Models191
A Feature-rich Compositional Embedding Model (FCM) for relation extraction that is expressive, generalizes to new domains, and is easy-to-implement that outperforms both previous compositional models and traditional feature rich models.
- Sentence Embedding Alignment for Lifelong Relation Extraction181
This work utilizes an explicit alignment model to mitigate the sentence embedding distortion of learned model when training on new data and new relations and significantly outperforms the state-of-the-art lifelong learning approaches.
- Improving Natural Language Inference Using External Knowledge in the Science Questions Domain178
A combination of techniques that harness knowledge graphs to improve performance on the NLI problem in the science questions domain and achieves the new state-of-the-art performance over the SciTail science questions dataset.
- Rethinking Cooperative Rationalization: Introspective Extraction and Complement Control163
An introspective model is introduced which explicitly predicts and incorporates the outcome into the selection process and is explicitly control the rationale complement via an adversary so as not to leave any useful information out of the selection.
- FaithDial : A Faithful Benchmark for Information-Seeking Dialogue159
It is found that the benefits of FaithDial generalize to zero-shot transfer on other datasets, such as CMU-Dog and TopicalChat, and human evaluation reveals that responses generated by models trained on FaithDial are perceived as more interpretable, cooperative, and engaging.
- 144
- 137
- Leveraging Sentence-level Information with Encoder LSTM for Semantic Slot Filling131
This paper enhances LSTM-based sequence labeling to explicitly model label dependencies and proposes another enhancement to incorporate the global information spanning over the whole input sequence to predict the label sequence.
- Improving Question Answering over Incomplete KBs with Knowledge-Aware Reader129
A new end-to-end question answering model, which learns to aggregate answer evidence from an incomplete knowledge base (KB) and a set of retrieved text snippets, which achieves consistent improvements across settings with different extents of KB incompleteness.
- Leveraging Abstract Meaning Representation for Knowledge Base Question Answering126
Neuro-Symbolic Question Answering (NSQA) is proposed, a modular KBQA system that leverages Abstract Meaning Representation (AMR) parses for task-independent question understanding and achieves state-of-the-art performance on two prominentKBQA datasets based on DBpedia.
- 119
- Deriving Machine Attention from Human Rationales115
It is demonstrated that even in the low-resource scenario, attention can be learned effectively, and this approach delivers significant gains over state-of-the-art baselines, yielding over 15% average error reduction on benchmark datasets.
- A Co-Matching Model for Multi-choice Reading Comprehension106
A new co-matching approach to this problem, which jointly models whether a passage can match both a question and a candidate answer, achieves state-of-the-art performance.
- Extracting Multiple-Relations in One-Pass with Pre-Trained Transformers104
This work focuses on the task of multiple relation extractions by encoding the paragraph only once, and builds the solution upon the pre-trained self-attentive models (Transformer), where it is shown that the approach is not only scalable but can also perform state-of-the-art on the standard benchmark ACE 2005.
- Learning Composition Models for Phrase Embeddings89
This work proposes efficient unsupervised and task-specific learning objectives that scale the model to large datasets and demonstrates improvements on both language modeling and several phrase semantic similarity tasks with various phrase lengths.
- TWEETQA: A Social Media Focused Question Answering Dataset87
This work presents the first large-scale dataset for QA over social media data, and shows that two recently proposed neural models that perform well on formal texts are limited in their performance when applied to this dataset.
- “Where Have All the Good Men Gone?” Gendered Interactions in Online Dating81
Gendered patterns of online dating and their implications for heterosexual union formation are explored and it is found that both men and women tend to send messages to the most socially desirable alters in the dating market, regardless of their own social desirability.
- 80
- Exploiting Rich Syntactic Information for Semantic Parsing with Graph-to-Sequence Model79
This paper first proposes to use the syntactic graph to represent three types of syntactic information, i.e., word order, dependency and constituency features; then employs a graph-to-sequence model to encode the Syntactic graph and decode a logical form.
- 75
- Accelerated Mini-batch Randomized Block Coordinate Descent Method.75
A mini-batch randomized block coordinate descent (MRBCD) method, which estimates the partial gradient of the selected block based on a mini-batch of randomly sampled data in each iteration, and shows that for strongly convex functions, the MRBCD method attains lower overall iteration complexity than existing RBCD methods.
- Query and Extract: Refining Event Extraction as Type-oriented Binary Decoding74
This work proposes a novel event extraction framework that uses event types and argument roles as natural language queries to extract candidate triggers and arguments from the input text and significantly outperforms existing methods on zero-shot event extraction.
- Out-of-Domain Detection for Low-Resource Text Classification Tasks66
Evaluations on real-world datasets show that the proposed solution outperforms state-of-the-art methods in zero-shot OOD detection task, while maintaining a competitive performance on ID classification task.
- A hybrid approach with optimization-based and metric-based meta-learner for few-shot learning66
This work proposes a hybrid meta-learning model called Meta-Metric-Learner which combines the merits of both optimization- and metric-based approaches and is able to handle flexible numbers of classes as well as generate more generalized metrics for classification across tasks.
- CASS65
This paper develops a generalizable chatbot architecture (CASS) to provide social support for community members in an online health community and contributes to fill the research gap on how a chatbot may influence the whole community's engagement.
- Educational Question Generation of Children Storybooks via Question Type Distribution Learning and Event-centric Summarization61
A novel question generation method that first learns the question type distribution of an input story paragraph, and then summarizes salient events which can be used to generate high-cognitive-demand questions.
- Self-Supervised Learning for Contextualized Extractive Summarization61
This paper aims to improve extractive summarization by introducing three auxiliary pre-training tasks that learn to capture the document-level context in a self-supervised fashion, and shows that after pre- training, a clean model with simple building blocks is able to outperform previous state-of-the-art that are carefully designed.
- Imposing Label-Relational Inductive Bias for Extremely Fine-Grained Entity Typing56
This work introduces a novel label-relational inductive bias, represented by a graph propagation layer that effectively encodes both global label co-occurrence statistics and word-level similarities and shows that a simple modification of the proposed graph layer can improve the performance on a conventional and widely-tested dataset that only includes KB-schema types.
- Multi-step Entity-centric Information Retrieval for Multi-Hop Question Answering56
This paper introduces an IR technique that uses information of entities present in the initially retrieved evidence to learn to ‘hop’ to other relevant evidence, which leads to significant boost in retrieval performance.
- Multilingual BERT Post-Pretraining Alignment54
This work proposes a simple method to align multilingual contextual embeddings as a post-pretraining step for improved cross-lingual transferability of the pretrained language models using parallel data.
- Combining Word Embeddings and Feature Embeddings for Fine-grained Relation Extraction54
A new model is proposed that conjoins features and word embeddings while maintaing a small number of parameters by learning feature embeddeddings jointly with the parameters of a compositional model, which attains state-of-the-art results on ACE and ERE fine-grained relation extraction.
- User Recommendations in Reciprocal and Bipartite Social Networks--An Online Dating Case Study47
A new collaborative-filtering model is proposed to improve user recommendations in bipartite and reciprocal social networks that considers a user's taste in picking others and attractiveness in being picked by others.
- Timeline Summarization based on Event Graph Compression via Time-Aware Optimal Transport45
A time-aware optimal transport distance is introduced for learning the compression model in an unsupervised manner and it is shown that the approach significantly improves on the state of the art on three real-world datasets, including two public standard benchmarks and the newly collected Timeline100 dataset.
- Leveraging Dependency Forest for Neural Medical Relation Extraction41
A graph neural network is used to represent the forests, automatically distinguishing the useful syntactic information from parsing noise, and shows that this method outperforms the standard tree-based methods, giving the state-of-the-art results in the literature.
- Argo: intelligent advertising by mining a user's interest from his photo collections40
The Argo system attempts to learn a user's profile from his shared photos and suggests relevant ads accordingly and represents ads in the topic space and matching their topic distributions with the target user interest.
- 38
- 38
- Hybrid Reinforcement Learning with Expert State Sequences36
This paper proposes a novel tensor-based model to infer the unobserved actions of the expert state sequences of an expert, while the expert actions are unobserved, and proposes a hybrid objective combining reinforcement learning and imitation learning.
- Large Language Models Can Self-Improve in Long-context Reasoning35
This work investigates the potential for LLMs to self-improve in long-context reasoning and proposes \the authors', an approach specifically designed for this purpose, which achieves superior performance compared to prior approaches that depend on data produced by human experts or advanced models.
- Interactive Fiction Game Playing as Multi-Paragraph Reading Comprehension with Reinforcement Learning35
This work takes a novel perspective of IF game solving and re-formulate it as Multi-Passage Reading Comprehension (MPRC) tasks and applies an object-centric historical observation retrieval strategy to mitigate the partial observability of the textual observations.
- 35
- Recommendation in Reciprocal and Bipartite Social Networks–A Case Study of Online Dating35
A new collaborative filtering model is proposed to improve user recommendations in reciprocal and bipartite social networks that considers a user's "taste" in picking others and "attractiveness" in being picked by others.
- Question Answering over Knowledge Bases by Leveraging Semantic Parsing and Neuro-Symbolic Reasoning.34
A semantic parsing and reasoning-based Neuro-Symbolic Question Answering system that achieves state-of-the-art performance on QALD-9 and LC-QuAD 1.0 and integrates multiple, reusable modules that are trained specifically for their individual tasks and do not require end-to-end training data.
- 33
- Group Chat Ecology in Enterprise Instant Messaging: How Employees Collaborate Through Multi-User Chat Channels on Slack31
This short paper aims to update the existing knowledge on how group chat is used in the context of today's organizations, and suggests design opportunities for the future of human-AI communication experience.
- Narrative Question Answering with Cutting-Edge Open-Domain QA Techniques: A Comprehensive Study31
The findings indicate that the event-centric questions dominate this task, which exemplifies the inability of existing QA models to handle event-oriented scenarios.
- Locally Training the Log-Linear Model for SMT31
This paper proposes a novel local training method, which significantly outperforms MERT with the maximal improvements up to 2.0 BLEU points, meanwhile its efficiency is comparable to that of the global method.
- Generalizable Resource Allocation in Stream Processing via Deep Reinforcement Learning29
This paper presents a graph-aware encoder-decoder framework to learn a generalizable resource allocation strategy that can properly distribute computation tasks of stream processing graphs unobserved from training data and proposes to leverage graph embedding to learn the structural information of the stream processing graph.
- Compound Embedding Features for Semi-supervised Learning29
This paper built compound features from continuous word embeddings based on clustering and showed that the compound features not only improved the performances on several NLP tasks, but also ran faster, suggesting the potential ofembeddings.
- Context-Aware Conversation Thread Detection in Multi-Party Chat28
This work proposes a novel Context-Aware Thread Detection (CATD) model that automatically disentangles intermingled conversation threads in chat logs and evaluates the model on four real-world datasets.
- Selection Bias Explorations and Debias Methods for Natural Language Sentence Matching Datasets27
This paper investigates the problem of selection bias on six NLSM datasets and finds that four out of them are significantly biased, and proposes a training and evaluation framework to alleviate the bias.
- Do Multi-hop Readers Dream of Reasoning Chains?27
The results demonstrate the existence of the potential improvement using explicit multi-hop reasoning and the necessity to develop models with better reasoning abilities and a co-matching-based method that leads to 13.1% error reduction with passage chains when applied to two of the base readers.
- Leveraging Sentence-level Information with Encoder LSTM for Natural Language Understanding.26
This paper enhances LSTM-based sequence labeling to explicitly model label dependencies and proposes another enhancement to incorporate the global information spanning over the whole input sequence to predict the label sequence.
- 25
- Interpretable Visual Reasoning via Induced Symbolic Space23
A new framework named object-centric compositional attention model (OCCAM) is designed to perform the visual reasoning task with object-level visual features, and a method to induce concepts of objects and relations using clues from the attention patterns between objects’ visual features and question words is proposed.
- 22
- 20
- 17
- 17
- Improving Reinforcement Learning Based Image Captioning with Natural Language Prior17
Quantitative and qualitative evaluations on benchmarks show that RL with the simple add-on module performs favorably against its counterpart in terms of both readability and speed of convergence.
- Coherency Improved Explainable Recommendation via Large Language Model16
This work proposes a novel framework that employs a large language model (LLM) to generate a rating, transforms it into a rating vector, and finally generates an explanation based on the rating vector and user-item information.
- Knowledge Base Relation Detection via Multi-View Matching16
This paper proposes a KB relation detection model via multi-view matching which utilizes more useful information extracted from question and KB which is designed in an end-to-end trainable neural network model.
- Diversifying landmark image search results by learning interested views from community photos16
A novel non-parametric TOC generation and set-based ranking algorithm, MoM-DPM Sets, is proposed as the key technology of Agate, which ranks landmark image search results considering their relevance, diversity and quality.
- 14
- 13
- Frustratingly Hard Evidence Retrieval for QA Over Books12
Besides achieving state-of-the-art on the NarrativeQA benchmark, this study reveals the difficulty of evidence retrieval in books with a wealth of experiments and analysis - which necessitates future effort on novel solutions forevidence retrieval in BookQA.
- An Empirical Study of Chinese Name Matching and Applications12
This work evaluates methods for name matching in Chinese, including both string matching and learning approaches, based on new representations for Chinese, which improves both name matching and a downstream entity clustering task.
- SitEmb-v1.5: Improved Context-Aware Dense Retrieval for Semantic Association and Long Story Comprehension11
It is shown that existing embedding models are not well-equipped to encode such situated context effectively, and thus a new training paradigm is introduced and the situated embedding models (SitEmb) are developed.
- 10
- 9
- 9
- 9
- 9
- 8
- 8
- Multi-Granular Text Encoding for Self-Explaining Categorization8
This work defines multi-granular ngrams as basic units for explanation, and organizes all n Grammars into a hierarchical structure, so that shorter n grams can be reused while computing longer n Grammar, and can extract intuitive multi- granular evidence to support its predictions.
- 8
- 7
- 7
- 6
- 6
- 6
- 6
- FaceID-6M: A Large-Scale, Open-Source FaceID Customization Dataset5
The first large-scale, open-source FaceID dataset containing 6 million high-quality text-image pairs is collected and released, and it is demonstrated that models trained on the FaceID-6M dataset achieve performance that is comparable to, and slightly better than currently available industrial models.
- Cross-lingual Projections between Languages from Different Families5
This work enhances the projection method based on word alignments by introducing target-language word representations as features and proposing a novel noise removing methodbased on these word representations.
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- 3
- 3
- 2
- 2
- 2
- An Energy-based Model for Word-level AutoCompletion in Computer-aided Translation2
This work proposes an energy-based model for WLAC, which enables the context hidden vector to capture crucial information from the source sentence and employs three simple yet effective strategies to put the model into practice.
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 1
- 1
- 1
- 1
- 1
- 1
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
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 2026-10-16. 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.