Jinggui Liang, Dung Vo, Yap Hong Xian, Hai Leong Chieu, Kian Ming A. Chai, Jing Jiang, Lizi Liao. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Yizhe Yang, Palakorn Achananuparp, Heyan Huang, Jing Jiang, Nicholas Gabriel Lim, Cameron Tan Shi Ern, Phey Ling Kit, Jenny Giam Xiuhui, John Pinto, Ee-Peng Lim. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Yizhe Yang, Palakorn Achananuparp, Heyan Huang, Jing Jiang, Phey Ling Kit, Nicholas Gabriel Lim, Cameron Tan Shi Ern, Ee-Peng Lim. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Humans possess a remarkable ability to interpret underspecified ambiguous statements by inferring their meanings from contexts such as visual inputs.This ability, however, may not be as developed in recent pre-trained visionlanguage models (VLMs).In this paper, we introduce a novel probing dataset called FO-CUS to evaluate whether state-of-the-art VLMs have this ability.FOCUS consists of underspecified sentences paired with image contexts and carefully designed probing questions.Our experiments reveal that VLMs still fall short in handling underspecification even when visual inputs that can help resolve the ambiguities are available.To further support research in underspecification, FOCUS will be released for public use.We hope this dataset will inspire further research on the reasoning and contextual understanding capabilities of VLMs.

Burak Satar, Zhixin Ma, Patrick Amadeus Irawan, Wilfried Ariel Mulyawan, Jing Jiang, Ee-Peng Lim, Chong-Wah Ngo. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

The data mixture for large language model pre-training significantly impacts performance, yet how to determine an effective mixture remains unclear. We propose RegMix to automatically identify a high-performing data mixture by formulating it as a regression task. RegMix trains many small models on diverse data mixtures, uses regression to predict performance of unseen mixtures, and applies the best predicted mixture to train a large-scale model with orders of magnitude more compute. To empirically validate RegMix, we train 512 models with 1M parameters for 1B tokens to fit the regression model and predict the best data mixture. Using this mixture we train a 1B parameter model for 25B tokens (i.e. 1000× larger and 25× longer) which we find performs best among 64 candidate 1B parameter models with other mixtures. Furthermore, RegMix consistently outperforms human selection in experiments involving models up to 7B models trained on 100B tokens, while matching or exceeding DoReMi using just 10% of the computational resources. Our experiments also show that (1) Data mixtures significantly impact performance; (2) Web corpora rather than data perceived as high-quality like Wikipedia have the strongest positive correlation with downstream performance; (3) Domains interact in complex ways often contradicting common sense, thus automatic approaches like RegMix are needed; (4) Data mixture effects transcend scaling laws. Our code is available at https://github.com/sail-sg/regmix.

Automatic LLM benchmarks, such as AlpacaEval 2.0, Arena-Hard-Auto, and MT-Bench, have become popular for evaluating language models due to their cost-effectiveness and scalability compared to human evaluation. Achieving high win rates on these benchmarks can significantly boost the promotional impact of newly released language models. This promotional benefit may motivate tricks, such as manipulating model output length or style to game win rates, even though several mechanisms have been developed to control length and disentangle style to reduce gameability. Nonetheless, we show that even a **"null model"** that always outputs a **constant** response (*irrelevant to input instructions*) can cheat automatic benchmarks and achieve top-ranked win rates: an $86.5\\%$ LC win rate on AlpacaEval 2.0; an $83.0$ score on Arena-Hard-Auto; and a $9.55$ score on MT-Bench. Moreover, the crafted cheating outputs are **transferable** because we assume that the instructions of these benchmarks (e.g., $805$ samples of AlpacaEval 2.0) are *private* and cannot be accessed. While our experiments are primarily proof-of-concept, an adversary could use LLMs to generate more imperceptible cheating responses, unethically benefiting from high win rates and promotional impact. Our findings call for the development of anti-cheating mechanisms for reliable automatic benchmarks. The code is available at https://github.com/sail-sg/Cheating-LLM-Benchmarks.

Text-to-Image (T2I) models have recently gained significant attention due to their ability to generate high-quality images and are consequently used in a wide range of applications. However, there are concerns about the gender bias of these models. Previous studies have shown that T2I models can perpetuate or even amplify gender stereotypes when provided with neutral text prompts (e.g., 'a photo of a CEO' is often associates with male images, while 'a photo of nurse' is often associates with female images). Researchers have proposed automated gender bias uncovering detectors for T2I models, but a crucial gap exists: no existing work comprehensively compares the various detectors and understands how the gender bias detected by them deviates from the actual situation. This study addresses this gap by validating previous gender bias detectors using a manually labeled dataset and comparing how the bias identified by various detectors deviates from the actual bias in T2I models, as verified by manual confirmation. We create a dataset consisting of 6,000 images generated from three cutting-edge T2I models, Stable Diffusion XL, Stable Diffusion 3, and Dreamlike Photoreal 2.0. During the human-labeling process, we find that all three T2I models generate a portion (12.48% on average) of low-quality images (e.g., generate images with no face present), where human annotators cannot determine the gender of the person. Our analysis reveals that all three T2I models show a preference for generating male images, with SDXL being the most biased. Additionally, images generated using prompts containing professional descriptions (e.g., lawyer or doctor) show the most bias. We evaluate seven gender bias detectors and find that none fully capture the actual level of bias in T2I models, with some detectors overestimating bias by up to 26.95%. We further investigate the causes of inaccurate estimations, highlighting the limitations of detectors in dealing with low-quality images. Based on our findings, we propose an enhanced detector called CLIP-Enhance, which most accurately measures the gender bias in T2I models, with a difference of only 0.47%-1.23%, and most effectively filters out 82.91% of low-quality images.1 We have made our dataset and code publicly available.

The recent success of large language models (LLMs) has attracted widespread interest to develop role-playing conversational agents personalized to the characteristics and styles of different speakers to enhance their abilities to perform both general and special purpose dialogue tasks.However, the ability to personalize the generated utterances to speakers, whether conducted by human or LLM, has not been well studied.To bridge this gap, our study introduces a novel evaluation challenge: speaker verification in agent-generated conversations, which aimed to verify whether two sets of utterances originate from the same speaker.To this end, we assemble a large dataset collection encompassing thousands of speakers and their utterances.We also develop and evaluate speaker verification models under experiment setups.We further utilize the speaker verification models to evaluate the personalization abilities of LLM-based role-playing models.Comprehensive experiments suggest that the current role-playing models fail in accurately mimicking speakers, primarily due to their inherent linguistic characteristics.

There are recent efforts to "personalize" large language models (LLMs) by assigning them specific personas.This paper explores the writing styles of such persona-assigned LLMs across different socio-demographic groups based on age, profession, location, and political affiliations, using three widely-used LLMs.Leveraging an existing style embedding model that produces detailed style attributes and latent Dirichlet allocation (LDA) for broad style analysis, we measure style differences using Kullback-Leibler divergence to compare LLM-generated and human-written texts.We find significant style differences among personas.This analysis emphasizes the need to consider socio-demographic factors in language modeling to accurately capture diverse writing styles used for communications.The findings also reveal the strengths and limitations of personalized LLMs, their potential uses, and the importance of addressing biases in their design.The code and data are available at: https://github.com/ ra-MANUJ-an/writing-style-persona Example from the training corpus That was the funniest thing so far this season.Sam SCREECHING and stabbin' wights all around in battle fury while more fall on him like throw pillows.Associated 20 Style Descriptors, ordered by score 'The author uses uncommon phrases.','The author uses descriptive words.','The author uses colorful language.','The author uses an energetic style.','The author uses a clever play on words.','The author is vivacious.','The author is using words to create a vivid and engaging atmosphere.','The author is using vivid descriptions.','The author is using punctuation to create a sense of tension and suspense.','The author is using male pronouns.','The author is intense in their writing.','The author is dramatic.','The author is captivating.','The author has a distinct and memorable style.','The author is creating a sense of anticipation and excitement.','The author is using a playful style.','The author is describing a current event.','The author uses victorious language.','The author is using a lighthearted tone.','The author uses singular subjects.'

Data attribution seeks to trace model outputs back to training data. With the recent development of diffusion models, data attribution has become a desired module to properly assign valuations for high-quality or copyrighted training samples, ensuring that data contributors are fairly compensated or credited. Several theoretically motivated methods have been proposed to implement data attribution, in an effort to improve the trade-off between computational scalability and effectiveness. In this work, we conduct extensive experiments and ablation studies on attributing diffusion models, specifically focusing on DDPMs trained on CIFAR-10 and CelebA, as well as a Stable Diffusion model LoRA-finetuned on ArtBench. Intriguingly, we report counter-intuitive observations that theoretically unjustified design choices for attribution empirically outperform previous baselines by a large margin, in terms of both linear datamodeling score and counterfactual evaluation. Our work presents a significantly more efficient approach for attributing diffusion models, while the unexpected findings suggest that at least in non-convex settings, constructions guided by theoretical assumptions may lead to inferior attribution performance. The code is available at https://github.com/sail-sg/D-TRAK.

Speculative decoding is a relatively new decoding framework that leverages small and efficient draft models to reduce the latency of LLMs. In this study, we introduce GliDe and CaPE, two low-hassle modifications to vanilla speculative decoding to further improve the decoding speed of a frozen LLM. Specifically, GliDe is a modified draft model architecture that reuses the cached keys and values from the target LLM, while CaPE is a proposal expansion method that uses the draft model’s confidence scores to help select additional candidate tokens for verification. Extensive experiments on different benchmarks demonstrate that our proposed GliDe draft model significantly reduces the expected decoding latency. Additional evaluation using walltime reveals that GliDe can accelerate Vicuna models up to 2.17x and further extend the improvement to 2.61x with CaPE. We will release our code, data, and the trained draft models.

A multimodal large language model (MLLM) agent can receive instructions, capture images, retrieve histories from memory, and decide which tools to use. Nonetheless, red-teaming efforts have revealed that adversarial images/prompts can jailbreak an MLLM and cause unaligned behaviors. In this work, we report an even more severe safety issue in multi-agent environments, referred to as infectious jailbreak. It entails the adversary simply jailbreaking a single agent, and without any further intervention from the adversary, (almost) all agents will become infected exponentially fast and exhibit harmful behaviors. To validate the feasibility of infectious jailbreak, we simulate multi-agent environments containing up to one million LLaVA-1.5 agents, and employ randomized pair-wise chat as a proof-of-concept instantiation for multi-agent interaction. Our results show that feeding an (infectious) adversarial image into the memory of any randomly chosen agent is sufficient to achieve infectious jailbreak. Finally, we derive a simple principle for determining whether a defense mechanism can provably restrain the spread of infectious jailbreak, but how to design a practical defense that meets this principle remains an open question to investigate.

Jinggui Liang, Lizi Liao, Hao Fei, Bobo Li, Jing Jiang. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.

Recently, Anil et al. (2024) show that many-shot (up to hundreds of) demonstrations can jailbreak state-of-the-art LLMs by exploiting their long-context capability. Nevertheless, is it possible to use few-shot demonstrations to efficiently jailbreak LLMs within limited context sizes? While the vanilla few-shot jailbreaking may be inefficient, we propose improved techniques such as injecting special system tokens like [/INST] and employing demo-level random search from a collected demo pool. These simple techniques result in surprisingly effective jailbreaking against aligned LLMs (even with advanced defenses). For example, our method achieves >80% (mostly >95%) ASRs on Llama-2-7B and Llama-3-8B without multiple restarts, even if the models are enhanced by strong defenses such as perplexity detection and/or SmoothLLM, which is challenging for suffix-based jailbreaking. In addition, we conduct comprehensive and elaborate (e.g., making sure to use correct system prompts) evaluations against other aligned LLMs and advanced defenses, where our method consistently achieves nearly 100% ASRs. Our code is available at https://github.com/sail-sg/I-FSJ.

In this paper, we address the challenge of detecting hateful memes in the low-resource setting where only a few labeled examples are available. Our approach leverages the compositionality of Low-rank adaptation (LoRA), a widely used parameter-efficient tuning technique. We commence by fine-tuning large language models (LLMs) with LoRA on selected tasks pertinent to hateful meme detection, thereby generating a suite of LoRA modules. These modules are capable of essential reasoning skills for hateful meme detection. We then use the few available annotated samples to train a module composer, which assigns weights to the LoRA modules based on their relevance. The model's learnable parameters are directly proportional to the number of LoRA modules. This modularized network, underpinned by LLMs and augmented with LoRA modules, exhibits enhanced generalization in the context of hateful meme detection. Our evaluation spans three datasets designed for hateful meme detection in a few-shot learning context. The proposed method demonstrates superior performance to traditional in-context learning, which is also more computationally intensive during inference.

Hateful meme detection is a challenging multimodal task that requires comprehension of both vision and language, as well as cross-modal interactions. Recent studies have tried to fine-tune pre-trained vision-language models (PVLMs) for this task. However, with increasing model sizes, it becomes important to leverage powerful PVLMs more efficiently, rather than simply fine-tuning them. Recently, researchers have attempted to convert meme images into textual captions and prompt language models for predictions. This approach has shown good performance but suffers from non-informative image captions. Considering the two factors mentioned above, we propose a probing-based captioning approach to leverage PVLMs in a zero-shot visual question answering (VQA) manner. Specifically, we prompt a frozen PVLM by asking hateful content-related questions and use the answers as image captions (which we call Pro-Cap), so that the captions contain information critical for hateful content detection. The good performance of models with Pro-Cap on three benchmarks validates the effectiveness and generalization of the proposed method1.

Knowledge base question answering (KBQA) aims to answer a question over a knowledge base (KB). Early studies mainly focused on answering simple questions over KBs and achieved great success. However, their performances on complex questions are still far from satisfactory. Therefore, in recent years, researchers propose a large number of novel methods, which looked into the challenges of answering complex questions. In this survey, we review recent advances in KBQA with the focus on solving complex questions, which usually contain multiple subjects, express compound relations, or involve numerical operations. In detail, we begin with introducing the complex KBQA task and relevant background. Then, we present two mainstream categories of methods for complex KBQA, namely semantic parsing-based (SP-based) methods and information retrieval-based (IR-based) methods. Specifically, we illustrate their procedures with flow designs and discuss their difference and similarity. Next, we summarize the challenges that these two categories of methods encounter when answering complex questions, and explicate advanced solutions as well as techniques used in existing work. After that, we discuss the potential impact of pre-trained language models (PLMs) on complex KBQA. To help readers catch up with SOTA methods, we also provide a comprehensive evaluation and resource about complex KBQA task. Finally, we conclude and discuss several promising directions related to complex KBQA for future research.

Minghuan Tan, Yong Dai, Duyu Tang, Zhangyin Feng, Guoping Huang, Jing Jiang, Jiwei Li, Shuming Shi. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2022.

A recent study by Feldman (2020) proposed a long-tail theory to explain the memorization behavior of deep learning models.However, memorization has not been empirically verified in the context of NLP, a gap addressed by this work.In this paper, we use three different NLP tasks to check if the long-tail theory holds.Our experiments demonstrate that top-ranked memorized training instances are likely atypical, and removing the top-memorized training instances leads to a more serious drop in test accuracy compared with removing training instances randomly.Furthermore, we develop an attribution method to better understand why a training instance is memorized.We empirically show that our memorization attribution method is faithful and share our interesting finding that the top-memorized parts of a training instance tend to be features negatively correlated with the class label.

Hateful meme classification is a challenging multimodal task that requires complex reasoning and contextual background knowledge.Ideally, we could leverage an explicit external knowledge base to supplement contextual and cultural information in hateful memes.However, there is no known explicit external knowledge base that could provide such hate speech contextual information.To address this gap, we propose PromptHate, a simple yet effective prompt-based model that prompts pre-trained language models (PLMs) for hateful meme classification.Specifically, we construct simple prompts and provide a few in-context examples to exploit the implicit knowledge in the pretrained RoBERTa language model for hateful meme classification.We conduct extensive experiments on two publicly available hateful and offensive meme datasets.Our experimental results show that PromptHate is able to achieve a high AUC of 90.96, outperforming state-ofthe-art baselines on the hateful meme classification task.We also perform fine-grained analyses and case studies on various prompt settings and demonstrate the effectiveness of the prompts on hateful meme classification.

Out-of-distribution (OOD) settings are used to measure a model's performance when the distribution of the test data is different from that of the training data.NLU models are known to suffer in OOD settings (Utama et al., 2020b).We study this issue from the perspective of causality, which sees confounding bias as the reason for models to learn spurious correlations.While a common solution is to perform intervention, existing methods handle only known and single confounder (Pearl and Mackenzie, 2018), but in many NLU tasks the confounders can be both unknown and multifactorial.In this paper, we propose a novel interventional training method called Bottom-up Automatic Intervention (BAI) that performs multi-granular intervention with identified multifactorial confounders.Our experiments on three NLU tasks, namely, natural language inference, fact verification and paraphrase identification, show the effectiveness of BAI for tackling different OOD settings.1

Multiple-Choice Question Answering (MCQA) is one of the challenging tasks in machine reading comprehension. The main challenge in MCQA is to extract "evidence" from the given context that supports the correct answer. In OpenbookQA dataset [1], the requirement of extracting "evidence" is particularly important due to the mutual independence of sentences in the context. Existing work tackles this problem by annotated evidence or distant supervision with rules which overly rely on human efforts. To address the challenge, we propose a simple yet effective approach termed evidence filtering to model the relationships between the encoded contexts with respect to different options collectively, and to potentially highlight the evidence sentences and filter out unrelated sentences. In addition to the effective reduction of human efforts of our approach compared, through extensive experiments on OpenbookQA, we show that the proposed approach outperforms the models that use the same backbone and more training data; and our parameter analysis also demonstrates the interpretability of our approach.

We design and evaluate VICTOR, an easy-to-apply module on top of a recommender system to mitigate misinformation. VICTOR takes an elegant, implicit approach to deliver fake-news verifications, such that readers of fake news can continuously access more verified news articles about fake-news events without explicit correction. We frame fake-news intervention within VICTOR as a graph-based question-answering (QA) task, with Q as a fake-news article and A as the corresponding verified articles. Specifically, VICTOR adopts reinforcement learning: it first considers fake-news readers’ preferences supported by underlying news recommender systems and then directs their reading sequence towards the verified news articles. To verify the performance of VICTOR, we collect and organize VERI, a new dataset consisting of real-news articles, user browsing logs, and fake-real news pairs for a large number of misinformation events. We evaluate zero-shot and few-shot VICTOR on VERI to simulate the never-exposed-ever and seen-before conditions of users while reading a piece of fake news. Results demonstrate that compared to baselines, VICTOR proactively delivers 6% more verified articles with a diversity increase of 7.5% to over 68% of at-risk users who have been exposed to fake news. Moreover, we conduct a field user study in which 165 participants evaluated fake news articles. Participants in the VICTOR condition show better exposure rates, proposal rates, and click rates on verified news articles than those in the other two conditions. Altogether, our work demonstrates the potentials of VICTOR, i.e., combat fake news by delivering verified information implicitly.

Yunshi Lan, Jing Jiang. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Sicheng Yu, Hao Zhang, Yulei Niu, Qianru Sun, Jing Jiang. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

There has been much interest in rumor detection using deep learning models in recent years.A well-known limitation of deep learning models is that they tend to learn superficial patterns, which restricts their generalization ability.We find that this is also true for cross-topic rumor detection.In this paper, we propose a method inspired by the "mixture of experts" paradigm.We assume that the prediction of the rumor class label given an instance is dependent on the topic distribution of the instance.After deriving a vector representation for each topic, given an instance, we derive a "topic mixture" vector for the instance based on its topic distribution.This topic mixture is combined with the vector representation of the instance itself to make rumor predictions.Our experiments show that our proposed method can outperform two baseline debiasing methods in a cross-topic setting.In a synthetic setting when we removed topic-specific words, our method also works better than the baselines, showing that our method does not rely on superficial features.

Knowledge base question answering (KBQA) aims to answer a question over a knowledge base (KB). Recently, a large number of studies focus on semantically or syntactically complicated questions. In this paper, we elaborately summarize the typical challenges and solutions for complex KBQA. We begin with introducing the background about the KBQA task. Next, we present the two mainstream categories of methods for complex KBQA, namely semantic parsing-based (SP-based) methods and information retrieval-based (IR-based) methods. We then review the advanced methods comprehensively from the perspective of the two categories. Specifically, we explicate their solutions to the typical challenges. Finally, we conclude and discuss some promising directions for future research.

Hateful and offensive content detection has been extensively explored in a single modality such as text. However, such toxic information could also be communicated via multimodal content such as online memes. Therefore, detecting multimodal hateful content has recently garnered much attention in academic and industry research communities. This paper aims to contribute to this emerging research topic by proposing DisMultiHate, which is a novel framework that performed the classification of multimodal hateful content. Specifically, DisMultiHate is designed to disentangle target entities in multimodal memes to improve the hateful content classification and explainability. We conduct extensive experiments on two publicly available hateful and offensive memes datasets. Our experiment results show that DisMultiHate is able to outperform state-of-the-art unimodal and multimodal baselines in the hateful meme classification task. Empirical case studies were also conducted to demonstrate DisMultiHate's ability to disentangle target entities in memes and ultimately showcase DisMultiHate's explainability of the multimodal hateful content classification task.

Multi-hop Knowledge Base Question Answering (KBQA) aims to find the answer entities that are multiple hops away in the Knowl- edge Base (KB) from the entities in the question. A major challenge is the lack of supervision signals at intermediate steps. Therefore, multi-hop KBQA algorithms can only receive the feedback from the final answer, which makes the learning unstable or ineffective. To address this challenge, we propose a novel teacher-student approach for the multi-hop KBQA task. In our approach, the stu- dent network aims to find the correct answer to the query, while the teacher network tries to learn intermediate supervision signals for improving the reasoning capacity of the student network. The major novelty lies in the design of the teacher network, where we utilize both forward and backward reasoning to enhance the learning of intermediate entity distributions. By considering bidi- rectional reasoning, the teacher network can produce more reliable intermediate supervision signals, which can alleviate the issue of spurious reasoning. Extensive experiments on three benchmark datasets have demonstrated the effectiveness of our approach on the KBQA task.

To address the increasingly significant issue of fake news, we develop a news reading platform in which we propose an implicit approach to reduce people's belief in fake news. Specifically, we leverage reinforcement learning to learn an intervention module on top of a recommender system (RS) such that the module is activated to replace RS to recommend news toward the verification once users touch the fake news. To examine the effect of the proposed method, we conduct a comprehensive evaluation with 89 human subjects and check the effective rate of change in belief but without their other limitations. Moreover, 84% participants indicate the proposed platform can help them defeat fake news. The demo video is available on YouTube https://youtu.be/wKI6nuXu_SM.

In Chinese, Chengyu are fixed phrases consisting of four characters. As a type of idioms, their meanings usually cannot be derived from their component characters. In this article, we study the task of recommending a Chengyu given a textual context. Observing some of the limitations with existing work, we propose a two-stage model, where during the first stage we re-train a Chinese BERT model by masking out Chengyu from a large Chinese corpus with a wide coverage of Chengyu. During the second stage, we fine-tune the re-trained, Chengyu-oriented BERT on a specific Chengyu recommendation dataset. We evaluate this method on ChID and CCT datasets and find that it can achieve the state of the art on both datasets. Ablation studies show that both stages of training are critical for the performance gain.

We address rumor detection by learning to differentiate between the community's response to real and fake claims in microblogs. Existing state-of-the-art models are based on tree models that model conversational trees. However, in social media, a user posting a reply might be replying to the entire thread rather than to a specific user. We propose a post-level attention model (PLAN) to model long distance interactions between tweets with the multi-head attention mechanism in a transformer network. We investigated variants of this model: (1) a structure aware self-attention model (StA-PLAN) that incorporates tree structure information in the transformer network, and (2) a hierarchical token and post-level attention model (StA-HiTPLAN) that learns a sentence representation with token-level self-attention. To the best of our knowledge, we are the first to evaluate our models on two rumor detection data sets: the PHEME data set as well as the Twitter15 and Twitter16 data sets. We show that our best models outperform current state-of-the-art models for both data sets. Moreover, the attention mechanism allows us to explain rumor detection predictions at both token-level and post-level.

Previous work on answering complex questions from knowledge bases usually separately addresses two types of complexity: questions with constraints and questions with multiple hops of relations.In this paper, we handle both types of complexity at the same time.Motivated by the observation that early incorporation of constraints into query graphs can more effectively prune the search space, we propose a modified staged query graph generation method with more flexible ways to generate query graphs.Our experiments clearly show that our method achieves the state of the art on three benchmark KBQA datasets.

In this paper, we study Multimodal Named Entity Recognition (MNER) for social media posts.Existing approaches for MNER mainly suffer from two drawbacks: (1) despite generating word-aware visual representations, their word representations are insensitive to the visual context; (2) most of them ignore the bias brought by the visual context.To tackle the first issue, we propose a multimodal interaction module to obtain both image-aware word representations and word-aware visual representations.To alleviate the visual bias, we further propose to leverage purely text-based entity span detection as an auxiliary module, and design a Unified Multimodal Transformer to guide the final predictions with the entity span predictions.Experiments show that our unified approach achieves the new state-of-the-art performance on two benchmark datasets.

Chinese idioms are special fixed phrases usually derived from ancient stories, whose meanings are oftentimes highly idiomatic and non-compositional.The Chinese idiom prediction task is to select the correct idiom from a set of candidate idioms given a context with a blank.We propose a BERT-based dual embedding model to encode the contextual words as well as to learn dual embeddings of the idioms.Specifically, we first match the embedding of each candidate idiom with the hidden representation corresponding to the blank in the context.We then match the embedding of each candidate idiom with the hidden representations of all the tokens in the context thorough context pooling.We further propose to use two separate idiom embeddings for the two kinds of matching.Experiments on a recently released Chinese idiom cloze test dataset show that our proposed method performs better than the existing state of the art.Ablation experiments also show that both context pooling and dual embedding contribute to the improvement of performance.

In this paper, we propose Cross-Thought, a novel approach to pre-training sequence encoder, which is instrumental in building reusable sequence embeddings for large-scale NLP tasks such as question answering.Instead of using the original signals of full sentences, we train a Transformer-based sequence encoder over a large set of short sequences, which allows the model to automatically select the most useful information for predicting masked words.Experiments on question answering and textual entailment tasks demonstrate that our pre-trained encoder can outperform state-of-the-art encoders trained with continuous sentence signals as well as traditional masked language modeling baselines.Our proposed approach also achieves new state of the art on HotpotQA (full-wiki setting) by improving intermediate information retrieval performance.1

The prevalent use of social media enables rapid spread of rumors on a massive scale, which leads to the emerging need of automatic rumor verification (RV). A number of previous studies focus on leveraging stance classification to enhance RV with multi-task learning (MTL) methods. However, most of these methods failed to employ pre-trained contextualized embeddings such as BERT, and did not exploit inter-task dependencies by using predicted stance labels to improve the RV task. Therefore, in this paper, to extend BERT to obtain thread representations, we first propose a Hierarchical Transformer, which divides each long thread into shorter subthreads, and employs BERT to separately represent each subthread, followed by a global Transformer layer to encode all the subthreads. We further propose a Coupled Transformer Module to capture the inter-task interactions and a Post-Level Attention layer to use the predicted stance labels for RV, respectively. Experiments on two benchmark datasets show the superiority of our Coupled Hierarchical Transformer model over existing MTL approaches.

Entity-level (aka target-dependent) sentiment analysis of social media posts has recently attracted increasing attention, and its goal is to predict the sentiment orientations over individual target entities mentioned in users' posts. Most existing approaches to this task primarily rely on the textual content, but fail to consider the other important data sources (e.g., images, videos, and user profiles), which can potentially enhance these text-based approaches. Motivated by the observation, we study entity-level multimodal sentiment classification in this article, and aim to explore the usefulness of images for entity-level sentiment detection in social media posts. Specifically, we propose an Entity-Sensitive Attention and Fusion Network (ESAFN) for this task. First, to capture the intra-modality dynamics, ESAFN leverages an effective attention mechanism to generate entity-sensitive textual representations, followed by aggregating them with a textual fusion layer. Next, ESAFN learns the entity-sensitive visual representation with an entity-oriented visual attention mechanism, followed by a gated mechanism to eliminate the noisy visual context. Moreover, to capture the inter-modality dynamics, ESAFN further fuses the textual and visual representations with a bilinear interaction layer. To evaluate the effectiveness of ESAFN, we manually annotate the sentiment orientation over each given entity based on two recently released multimodal NER datasets, and show that ESAFN can significantly outperform several highly competitive unimodal and multimodal methods.

In this paper, we study abstractive review summarization. Observing that review summaries often consist of aspect words, opinion words and context words, we propose a two-stage reinforcement learning approach, which first predicts the output word type from the three types, and then leverages the predicted word type to generate the final word distribution. Experimental results on two Amazon product review datasets demonstrate that our method can consistently outperform several strong baseline approaches based on ROUGE scores.

Knowledge Base Question Answering (KBQA) has attracted much attention and recently there has been more interest in multi-hop KBQA. In this paper, we propose a novel iterative sequence matching model to address several limitations of previous methods for multi-hop KBQA. Our method iteratively grows the candidate relation paths that may lead to answer entities. The method prunes away less relevant branches and incrementally assigns matching scores to the paths. Empirical results demonstrate that our method can significantly outperform existing methods on three different benchmark datasets.

Knowledge base question answering (KBQA) is an important task in natural language processing. Existing methods for KBQA usually start with entity linking, which considers mostly named entities found in a question as the starting points in the KB to search for answers to the question. However, relying only on entity linking to look for answer candidates may not be sufficient. In this paper, we propose to perform topic unit linking where topic units cover a wider range of units of a KB. We use a generation-and-scoring approach to gradually refine the set of topic units. Furthermore, we use reinforcement learning to jointly learn the parameters for topic unit linking and answer candidate ranking in an end-to-end manner. Experiments on three commonly used benchmark datasets show that our method consistently works well and outperforms the previous state of the art on two datasets.

As an important task in Sentiment Analysis, Target-oriented Sentiment Classification (TSC) aims to identify sentiment polarities over each opinion target in a sentence. However, existing approaches to this task primarily rely on the textual content, but ignoring the other increasingly popular multimodal data sources (e.g., images), which can enhance the robustness of these text-based models. Motivated by this observation and inspired by the recently proposed BERT architecture, we study Target-oriented Multimodal Sentiment Classification (TMSC) and propose a multimodal BERT architecture. To model intra-modality dynamics, we first apply BERT to obtain target-sensitive textual representations. We then borrow the idea from self-attention and design a target attention mechanism to perform target-image matching to derive target-sensitive visual representations. To model inter-modality dynamics, we further propose to stack a set of self-attention layers to capture multimodal interactions. Experimental results show that our model can outperform several highly competitive approaches for TSC and TMSC.

Making use of knowledge bases to answer questions (KBQA) is a key direction in question answering systems. Researchers have developed a diverse range of methods to address this problem, but there are still some limitations with the existing methods. Specifically, the existing neural network-based methods for KBQA have not taken advantage of the recent “matching-aggregation” framework for the sequence matching, and when representing a candidate answer entity, they may not choose the most useful context of the candidate for matching. In this paper, we explore the use of a “matching-aggregation” framework to match candidate answers with questions. We further make use of question-specific contextual relations to enhance the representations of candidate answer entities. Our complete method is able to achieve state-of-the-art performance on two benchmark datasets: WebQuestions and SimpleQuestions.

Extracting aspect terms and opinion terms are two fundamental tasks in opinion mining. The recent success of deep learning has inspired various neural network architectures, which have been shown to achieve highly competitive performance in these two tasks. However, most existing methods fail to explicitly consider the syntactic relations among aspect terms and opinion terms, which may lead to the inconsistencies between the model predictions and the syntactic constraints. To this end, we first apply a multi-task learning framework to implicitly capture the relations between the two tasks, and then propose a global inference method by explicitly modelling several syntactic constraints among aspect term extraction and opinion term extraction to uncover their intra-task and inter-task relationship, which seeks an optimal solution over the neural predictions for both tasks. Extensive evaluations on three benchmark datasets demonstrate that our global inference approach is able to bring consistent improvements over several base models in different scenarios.

In recent years researchers have achieved considerable success applying neural network methods to question answering (QA). These approaches have achieved state of the art results in simplified closed-domain settings such as the SQuAD (Rajpurkar et al. 2016) dataset, which provides a pre-selected passage, from which the answer to a given question may be extracted. More recently, researchers have begun to tackle open-domain QA, in which the model is given a question and access to a large corpus (e.g., wikipedia) instead of a pre-selected passage (Chen et al. 2017a). This setting is more complex as it requires large-scale search for relevant passages by an information retrieval component, combined with a reading comprehension model that “reads” the passages to generate an answer to the question. Performance in this setting lags well behind closed-domain performance. In this paper, we present a novel open-domain QA system called Reinforced Ranker-Reader (R3), based on two algorithmic innovations. First, we propose 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. Second, we propose a novel method that jointly trains the Ranker along with an answer-extraction Reader model, based on reinforcement learning. We report extensive experimental results showing that our method significantly improves on the state of the art for multiple open-domain QA datasets.

In this paper, we study how we can improve a deep learning approach to textual entailment by incorporating lexical entailment relations from WordNet. Our idea is to embed the lexical entailment knowledge contained in WordNet in specially-learned word vectors, which we call “entailment vectors.” We present a standard neural network model and a novel set-theoretic model to learn these entailment vectors from word pairs with known lexical entailment relations derived from WordNet. We further incorporate these entailment vectors into a decomposable attention model for textual entailment and evaluate the model on the SICK and the SNLI dataset. We find that using these special entailment word vectors, we can significantly improve the performance of textual entailment compared with a baseline that uses only standard word2vec vectors. The final performance of our model is close to or above the state of the art, but our method does not rely on any manually-crafted rules or extensive syntactic features.

In this paper, we target at improving the performance of multi-label emotion classification with the help of sentiment classification.Specifically, we propose a new transfer learning architecture to divide the sentence representation into two different feature spaces, which are expected to respectively capture the general sentiment words and the other important emotion-specific words via a dual attention mechanism.Extensive experimental results demonstrate that our transfer learning approach can outperform several strong baselines and achieve the state-of-the-art performance on two benchmark datasets.

Nowadays, it is a heated topic for many industries to build automatic question-answering (QA) systems. A key solution to these QA systems is to retrieve from a QA knowledge base the most similar question of a given question, which can be reformulated as a paraphrase identification (PI) or a natural language inference (NLI) problem. However, most existing models for PI and NLI have at least two problems: They rely on a large amount of labeled data, which is not always available in real scenarios, and they may not be efficient for industrial applications. In this paper, we study transfer learning for the PI and NLI problems, aiming to propose a general framework, which can effectively and efficiently adapt the shared knowledge learned from a resource-rich source domain to a resource-poor target domain. Specifically, since most existing transfer learning methods only focus on learning a shared feature space across domains while ignoring the relationship between the source and target domains, we propose to simultaneously learn shared representations and domain relationships in a unified framework. Furthermore, we propose an efficient and effective hybrid model by combining a sentence encoding-based method and a sentence interaction-based method as our base model. Extensive experiments on both paraphrase identification and natural language inference demonstrate that our base model is efficient and has promising performance compared to the competing models, and our transfer learning method can help to significantly boost the performance. Further analysis shows that the inter-domain and intra-domain relationship captured by our model are insightful. Last but not least, we deploy our transfer learning model for PI into our online chatbot system, which can bring in significant improvements over our existing system. Finally, we launch our new system on the chatbot platform Eva in our E-commerce site AliExpress.

Opinion target extraction is a fundamental task in opinion mining. In recent years, neural network based supervised learning methods have achieved competitive performance on this task. However, as with any supervised learning method, neural network based methods for this task cannot work well when the training data comes from a different domain than the test data. On the other hand, some rule-based unsupervised methods have shown to be robust when applied to different domains. In this work, we use rule-based unsupervised methods to create auxiliary labels and use neural network models to learn a hidden representation that works well for different domains. When this hidden representation is used for opinion target extraction, we find that it can outperform a number of strong baselines with a large margin.

Liangguo Wang, Jing Jiang, Hai Leong Chieu, Chen Hui Ong, Dandan Song, Lejian Liao. Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2017.

In this paper, we study domain adaptation with a state-of-the-art hierarchical neural network for document-level sentiment classification. We first design a new auxiliary task based on sentiment scores of domain-independent words. We then propose two neural network architectures to respectively induce document embeddings and sentence embeddings that work well for different domains. When these document and sentence embeddings are used for sentiment classification, we find that with both pseudo and external sentiment lexicons, our proposed methods can perform similarly to or better than several highly competitive domain adaptation methods on a benchmark dataset of product reviews.

Microblogging services such as Sina Weibo and Twitter allow users to create tags explicitly indicated by the # symbol. In Sina Weibo, these tags are called microtopics , and in Twitter, they are called hashtags . In Sina Weibo, each microtopic has a designate page and can be directly visited or commented on. Recommending these microtopics to users based on their interests can help users efficiently acquire information. However, it is non-trivial to recommend microtopics to users to satisfy their information needs. In this article, we investigate the task of personalized microtopic recommendation, which exhibits two challenges. First, users usually do not give explicit ratings to microtopics. Second, there exists rich information about users and microtopics, for example, users' published content and biographical information, but it is not clear how to best utilize such information. To address the above two challenges, we propose a joint probabilistic latent factor model to integrate rich information into a matrix factorization-based solution to microtopic recommendation. Our model builds on top of collaborative filtering, content analysis, and feature regression. Using two real-world datasets, we evaluate our model with different kinds of content and contextual information. Experimental results show that our model significantly outperforms a few competitive baseline methods, especially in the circumstance where users have few adoption behaviors.

Microblogging services allow users to create hashtags to categorize their posts. In recent years, the task of recommending hashtags for microblogs has been given increasing attention. However, most of existing methods depend on hand-crafted features. Motivated by the successful use of long short-term memory (LSTM) for many natural language processing tasks, in this paper, we adopt LSTM to learn the representation of a microblog post. Observing that hashtags indicate the primary topics of microblog posts, we propose a novel attention-based LSTM model which incorporates topic modeling into the LSTM architecture through an attention mechanism. We evaluate our model using a large real-world dataset. Experimental results show that our model significantly outperforms various competitive baseline methods. Furthermore, the incorporation of topical attention mechanism gives more than 7.4% improvement in F1 score compared with standard LSTM method.

Relation classification is the task of classifying the semantic relations between entity pairs in text. Observing that existing work has not fully explored using different representations for relation instances, especially in order to better handle the asymmetry of relation types, in this paper, we propose a neural network based method for relation classification that combines the raw sequence and the shortest dependency path representations of relation instances and uses mirror instances to perform pairwise relation classification. We evaluate our proposed models on the SemEval-2010 Task 8 dataset. The empirical results show that with two additional features, our model achieves the state-of-the-art result of F1 score of 85.7.

National Research Foundation (NRF) Singapore under International Research Centres in Singapore Funding Initiative

Social networks are an important class of networks that span a wide variety of media, ranging from social websites such as Facebook and Google Plus, citation networks of academic papers and patents, caller networks in telecommunications, and hyperlinked document collections such as Wikipedia - to name a few. Many of these social networks now exceed millions of users or actors, each of which may be associated with rich attribute data such as user profiles in social websites and caller networks, or subject classifications in document collections and citation networks. Such attribute data is often incomplete for a number of reasons - for example, users may be unwilling to spend the effort to complete their profiles, while in the case of document collections, there may be insufficient human labor to accurately classify all documents. At the same time, the tie or link information in these networks may also be incomplete - in social websites, users may simply be unaware of potential acquaintances, while in citation networks, authors may be unaware of appropriate literature that should be referenced. Completing and predicting these missing attributes and ties is important to a spectrum of applications, such as recommendation, personalized search, and targeted advertising, yet large social networks can pose a scalability challenge to existing algorithms designed for this task. Towards this end, we propose an integrative probabilistic model, SLR, that captures both attribute and tie information simultaneously, and can be used for attribute completion and tie prediction, in order to enable the above mentioned applications. A key innovation in our model is the use of triangle motifs to represent ties in the network, in order to scale to networks with millions of nodes and beyond. Experiments on real world datasets show that SLR significantly improves the accuracy of attribute prediction and tie prediction compared to well-known methods, and our distributed, multi-machine implementation easily scales up to millions of users. In addition to fast and accurate attribute and tie prediction, we also demonstrate how SLR can identify the attributes most responsible for homophily within the network, thus revealing which attributes drive network tie formation.

Natural language inference (NLI) is a fundamentally important task in natural language processing that has many applications.The recently released Stanford Natural Language Inference (SNLI) corpus has made it possible to develop and evaluate learning-centered methods such as deep neural networks for natural language inference (NLI).In this paper, we propose a special long short-term memory (LSTM) architecture for NLI.Our model builds on top of a recently proposed neural attention model for NLI but is based on a significantly different idea.Instead of deriving sentence embeddings for the premise and the hypothesis to be used for classification, our solution uses a match-LSTM to perform wordby-word matching of the hypothesis with the premise.This LSTM is able to place more emphasis on important word-level matching results.In particular, we observe that this LSTM remembers important mismatches that are critical for predicting the contradiction or the neutral relationship label.On the SNLI corpus, our model achieves an accuracy of 86.1%, outperforming the state of the art.

Twitter has become one of the largest microblogging platforms for users around the world to share anything happening around them with friends and beyond. A bursty topic in Twitter is one that triggers a surge of relevant tweets within a short period of time, which often reflects important events of mass interest. How to leverage Twitter for early detection of bursty topics has therefore become an important research problem with immense practical value. Despite the wealth of research work on topic modelling and analysis in Twitter, it remains a challenge to detect bursty topics in real-time. As existing methods can hardly scale to handle the task with the tweet stream in real-time, we propose in this paper$\sf {TopicSketch}$, a sketch-based topic model together with a set of techniques to achieve real-time detection. We evaluate our solution on a tweet stream with over 30 million tweets. Our experiment results show both efficiency and effectiveness of our approach. Especially it is also demonstrated that$\sf {TopicSketch}$on a single machine can potentially handle hundreds of millions tweets per day, which is on the same scale of the total number of daily tweets in Twitter, and present bursty events in finer-granularity.

Online debate forums are important social media for people to voice their opinions and debate with each other. Mining user stances or viewpoints from these forums has been a popular research topic. However, most current work does not address an important problem: for a specific issue, there may not be many users participating and expressing their opinions. Despite the sparsity of user stances, users may provide rich side information; for example, users may write arguments to back up their stances, interact with each other, and provide biographical information. In this work, we propose an integrated model to leverage side information. Our proposed method is a regression-based latent factor model which jointly models user arguments, interactions, and attributes. Our method can perform stance prediction for both warm-start and cold-start users. We demonstrate in experiments that our method has promising results on both micro-level and macro-level stance prediction.

As the rapid growth of online social media attracts a large number of Internet users, the large volume of content generated by these users also provides us with an opportunity to study the lexical variation of people of different ages. In this paper, we present a latent variable model that jointly models the lexical content of tweets and Twitter users’ ages. Our model inherently assumes that a topic has not only a word distribution but also an age distribution. We propose a Gibbs-EM algorithm to perform inference on our model. Empirical evaluation shows that our model can learn meaningful age-specific topics such as “school” for teenagers and “health” for older people. Our model can also be used for age prediction and performs better than a number of baseline methods.

Threaded debate forums have become one of the major social media platforms. Usually people argue with one another using not only claims and evidences about the topic under discussion but also language used to organize them, which we refer to as shell. In this paper, we study how to separate shell from topical contents using unsupervised methods. Along this line, we develop a latent variable model named Shell Topic Model (STM) to jointly model both topics and shell. Experiments on real online debate data show that our model can find both meaningful shell and topics. The results also show the effectiveness of our model by comparing it with several baselines in shell phrases extraction and document modeling.

Recommendation and review sites offer a wealth of information beyond ratings. For instance, on IMDb users leave reviews, commenting on different aspects of a movie (e.g. actors, plot, visual effects), and expressing their sentiments (positive or negative) on these aspects in their reviews. This suggests that uncovering aspects and sentiments will allow us to gain a better understanding of users, movies, and the process involved in generating ratings.

A common problem in large urban cities is the huge number of retail options available. In response, a number of shopping assistance applications have been created for mobile phones. However, these applications mostly allow users to know where stores are or find promotions on specific items. What i

Due to the fast development of social media on the Web, Twitter has become one of the major platforms for people to express themselves. Because of the wide adoption of Twitter, events like breaking news and release of popular videos can easily catch people's attention and spread rapidly on Twitter, and the number of relevant tweets approximately reflects the impact of an event. Event identification and analysis on Twitter has thus become an important task. Recently the Recurrent Chinese Restaurant Process (RCRP) has been successfully used for event identification from news streams and news-centric social media streams. However, these models cannot be directly applied to Twitter based on our preliminary experiments mainly for two reasons: (1) Events emerge and die out fast on Twitter, while existing models ignore this burstiness property. (2) Most Twitter posts are personal interest oriented while only a small fraction is event related. Motivated by these challenges, we propose a new nonparametric model which considers burstiness. We further combine this model with traditional topic models to identify both events and topics simultaneously. Our quantitative evaluation provides sufficient evidence that our model can accurately detect meaningful events. Our qualitative evaluation also shows interesting analysis for events on Twitter.

Most social networking platforms protect user's private information by limiting access to it to a small group of members, typically friends of the user, while allowing (virtually) everyone's access to the user's public data. In this paper, we exploit public data available on Facebook to infer users' undisclosed interests on their profile pages. In particular, we infer their undisclosed interests from the public data fetched using Graph APIs provided by Facebook. We demonstrate that simply liking a Facebook page does not corroborate that the user is interested in the page. Instead, we perform sentiment-oriented mining on various attributes of a Facebook page to determine the user's real interests. Our experiments conducted on over 34,000 public pages collected from Facebook and data from volunteers show that our inference technique can infer interests that are often hidden by users on their personal profile with moderate accuracy. We are able to disclose 22 interests of a user and find more than 80,097 users with at least 2 interests. We also show how this inferred information can be used to break a preference based backup authentication system.

Online discussion forums are popular social media platforms for users to express their opinions and discuss controversial issues with each other. To automatically identify the sides/stances of posts or users from textual content in forums is an important task to help mine online opinions. To tackle the task, it is important to exploit user posts that implicitly contain support and dispute (interaction) information. The challenge we face is how to mine such interaction information from the content of posts and how to use them to help identify stances. This paper proposes a two-stage solution based on latent variable models: an interaction feature identification stage to mine interaction features from structured debate posts with known sides and reply intentions; and a clustering stage to incorporate interaction features and model the interplay between interactions and sides for debate side clustering. Empirical evaluation shows that the learned interaction features provide good insights into user interactions and that with these features our debate side model shows significant improvement over other baseline methods.

Community Question Answering (CQA) websites, where people share expertise on open platforms, have become large repositories of valuable knowledge. To bring the best value out of these knowledge repositories, it is critically important for CQA services to know how to find the right experts, retrieve archived similar questions and recommend best answers to new questions. To tackle this cluster of closely related problems in a principled approach, we proposed Topic Expertise Model (TEM), a novel probabilistic generative model with GMM hybrid, to jointly model topics and expertise by integrating textual content model and link structure analysis. Based on TEM results, we proposed CQARank to measure user interests and expertise score under different topics. Leveraging the question answering history based on long-term community reviews and voting, our method could find experts with both similar topical preference and high topical expertise. Experiments carried out on Stack Overflow data, the largest CQA focused on computer programming, show that our method achieves significant improvement over existing methods on multiple metrics.

With the rapid growth of social media, Twitter has become one of the most widely adopted platforms for people to post short and instant message.On the one hand, people tweets about their daily lives, and on the other hand, when major events happen, people also follow and tweet about them.Moreover, people's posting behaviors on events are often closely tied to their personal interests.In this paper, we try to model topics, events and users on Twitter in a unified way.We propose a model which combines an LDA-like topic model and the Recurrent Chinese Restaurant Process to capture topics and events.We further propose a duration-based regularization component to find bursty events.We also propose to use event-topic affinity vectors to model the association between events and topics.Our experiments shows that our model can accurately identify meaningful events and the event-topic affinity vectors are effective for event recommendation and grouping events by topics.

We explore Debatepedia, a communityauthored encyclopedia of sociopolitical debates, as evidence for inferring a lowdimensional, human-interpretable representation in the domain of issues and positions.We introduce a generative model positing latent topics and cross-cutting positions that gives special treatment to person mentions and opinion words.We evaluate the resulting representation's usefulness in attaching opinionated documents to arguments and its consistency with human judgments about positions.0.1 0.2 0.3 0.4 0.5 comment, minimum, wage, poverty, capitalism nuclear, weapons, iran, states, threat party, vote, republican, political, voters energy, gas, power, fuel, wind

Twitter has become one of the largest platforms for users around the world to share anything happening around them with friends and beyond. A bursty topic in Twitter is one that triggers a surge of relevant tweets within a short time, which often reflects important events of mass interest. How to leverage Twitter for early detection of bursty topics has therefore become an important research problem with immense practical value. Despite the wealth of research work on topic modeling and analysis in Twitter, it remains a huge challenge to detect bursty topics in real-time. As existing methods can hardly scale to handle the task with the tweet stream in real-time, we propose in this paper Topic Sketch, a novel sketch-based topic model together with a set of techniques to achieve real-time detection. We evaluate our solution on a tweet stream with over 30 million tweets. Our experiment results show both efficiency and effectiveness of our approach. Especially it is also demonstrated that Topic Sketch can potentially handle hundreds of millions tweets per day which is close to the total number of daily tweets in Twitter and present bursty event in finer-granularity.

Textual information exchanged among users on online social network platforms provides deep understanding into users’ interest and behavioral patterns. However, unlike traditional text-dominant settings such as offline publishing, one distinct feature for online social network is users’ rich interactions with the textual content, which, unfortunately, has not yet been well incorporated in the existing topic modeling frameworks. In this paper, we propose an LDA-based behavior-topic model (B-LDA) which jointly models user topic interests and behavioral patterns. We focus the study of the model on online social network settings such as microblogs like Twitter where the textual content is relatively short but user interactions on them are rich. We conduct experiments on real Twitter data to demonstrate that the topics obtained by our model are both informative and insightful. As an application of our B-LDA model, we also propose a Twitter followee recommendation algorithm combining B-LDA and LDA, which we show in a quantitative experiment outperforms LDA with a significant margin.

Socialness refers to the ability to elicit social interaction and social links among people. It is a concept often associated with individuals. Although there are tangible benefits in socialness, there is little research in its modeling. In this paper, we study socialness as a property that can be associated with items, beyond its traditional association with people. We aim to model an item's socialness as a quantitative measure based on the how popular the item is adopted by members of multiple communities. We propose two socialness models, namely Basic and Mutual Dependency, to compute item socialness based on different sets of principles. In developing the Mutual Dependency Model, we demonstrate that items' socialness can be related to the socialness of communities. Our model have been evaluated on a set of users and application items from a mobile social network. We also conducted experiments to study how socialness can be related to network effects such as homophily, social influence and friendship formation.

Online software forums provide a huge amount of valuable content. Developers and users often ask questions and receive answers from such forums. The availability of a vast amount of thread discussions in forums provides ample opportunities for knowledge acquisition and summarization. For a given search query, current search engines use traditional information retrieval approach to extract webpages containing relevant keywords. However, in software forums, often there are many threads containing similar keywords where each thread could contain a lot of posts as many as 1,000 or more. Manually finding relevant answers from these long threads is a painstaking task to the users. Finding relevant answers is particularly hard in software forums as: complexities of software systems cause a huge variety of issues often expressed in similar technical jargons, and software forum users are often expert internet users who often posts answers in multiple venues creating many duplicate posts, often without satisfying answers, in the world wide web. To address this problem, this paper provides a semantic search engine framework to process software threads and recover relevant answers according to user queries. Different from standard information retrieval engine, our framework infer semantic tags of posts in the software forum threads and utilize these tags to recover relevant answer posts. In our case study, we analyze 6,068 posts from three software forums. In terms of accuracy of our inferred tags, we could achieve on average an overall precision, recall and F-measure of 67%, 71%, and 69% respectively. To empirically study the benefit of our overall framework, we also conduct a user-assisted study which shows that as compared to a standard information retrieval approach, our proposed framework could increase mean average precision from 17% to 71% in retrieving relevant answers to various queries and achieve a Normalized Discounted Cumulative Gain (nDCG) @1 score of 91.2% and nDCG@2 score of 71.6%.

In email networks, user behaviors affect the way emails are sent and replied. While knowing these user behaviors can help to create more intelligent email services, there has not been much research into mining these behaviors. In this paper, we investigate user engagingness and responsiveness as two interaction behaviors that give us useful insights into how users email one another. Engaging users are those who can effectively solicit responses from other users. Responsive users are those who are willing to respond to other users. By modeling such behaviors, we are able to mine them and to identify engaging or responsive users. This paper proposes four types of models to quantify engagingness and responsiveness of users. These behaviors can be used as features in the email reply order prediction task which predicts the email reply order given an email pair. Our experiments show that engagingness and responsiveness behavior features are more useful than other non-behavior features in building a classifier for the email reply order prediction task. When combining behavior and non-behavior features, our classifier is also shown to predict the email reply order with good accuracy.

Bursty features in text streams are very useful in many text mining applications. Most existing studies detect bursty features based purely on term frequency changes without taking into account the semantic contexts of terms, and as a result the detected bursty features may not always be interesting or easy to interpret. In this paper we propose to model the contexts of bursty features using a language modeling approach. We then propose a novel topic diversity-based metric using the context models to find newsworthy bursty features. We also propose to use the context models to automatically assign meaningful tags to bursty features. Using a large corpus of a stream of news articles, we quantitatively show that the proposed context language models for bursty features can effectively help rank bursty features based on their newsworthiness and to assign meaningful tags to annotate bursty features.

Trust reciprocity, a special form of link reciprocity, exists in many networks of trust among users. In this paper, we seek to determine the extent to which reciprocity exists in a trust network and develop quantitative models for measuring reciprocity and reciprocity related behaviors. We identify several reciprocity behaviors and their respective measures. These behavior measures can be employed for predicting if a trustee will return trust to her trustor given that the latter initiates a trust link earlier. We develop for this reciprocal trust prediction task a number of ranking method and classification methods, and evaluated them on an Epinions trust network data. Our results show that reciprocity related behaviors provide good features for both ranking and classification based methods under different parameter settings.

This paper focuses on the problem of identifying influential users of micro-blogging services. Twitter, one of the most notable micro-blogging services, employs a social-networking model called "following", in which each user can choose who she wants to "follow" to receive tweets from without requiring the latter to give permission first. In a dataset prepared for this study, it is observed that (1) 72.4% of the users in Twitter follow more than 80% of their followers, and (2) 80.5% of the users have 80% of users they are following follow them back. Our study reveals that the presence of "reciprocity" can be explained by phenomenon of homophily. Based on this finding, TwitterRank, an extension of PageRank algorithm, is proposed to measure the influence of users in Twitter. TwitterRank measures the influence taking both the topical similarity between users and the link structure into account. Experimental results show that TwitterRank outperforms the one Twitter currently uses and other related algorithms, including the original PageRank and Topic-sensitive PageRank.

This paper analyzes the trustor and trustee factors that lead to inter-personal trust using a well studied Trust Antecedent framework in management science. To apply these factors to trust ranking problem in online rating systems, we derive features that correspond to each factor and develop different trust ranking models. The advantage of this approach is that features relevant to trust can be systematically derived so as to achieve good prediction accuracy. Through a series of experiments on real data from Epinions, we show that even a simple model using the derived features yields good accuracy and outperforms MoleTrust, a trust propagation based model. SVM classifiers using these features also show improvements.