Zhong-Zhi Li, Xiao Liang, Zihao Tang, Lei Ji, Peijie Wang, Haotian Xu, Xing W, Haizhen Huang, Weiwei Deng, Yeyun Gong, Zhijiang Guo, Xiao Liu, Fei Yin, Cheng-Lin Liu. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Kailai Yang, Xiao Liu, Lei Ji, Hao Li, Xiao Liang, Zhiwei Liu, Yeyun Gong, Peng Cheng, Mao Yang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

As large language models (LLMs) advance, their inability to autonomously execute tasks by directly interacting with external tools remains a critical limitation. Traditional methods rely on inputting tool descriptions as context, which is constrained by context length and requires separate, often inefficient, retrieval mechanisms. We introduce ToolGen, a paradigm shift that integrates tool knowledge directly into the LLM’s parameters by representing each tool as a unique token. This enables the LLM to generate tool calls and arguments as part of its next token prediction capabilities, seamlessly blending tool invocation with language generation. Our framework allows the LLM to access and utilize a vast amount of tools with no additional retrieval step, significantly enhancing both performance and scalability. Experimental results with over 47,000 tools show that ToolGen not only achieves superior results in both tool retrieval and autonomous task completion but also sets the stage for a new era of AI agents that can adapt to tools across diverse domains. By fundamentally transforming tool retrieval into a generative process, ToolGen paves the way for more versatile, efficient, and autonomous AI systems. ToolGen enables end-to-end tool learning and opens opportunities for integration with other advanced techniques such as chain-of-thought and reinforcement learning, thereby expanding the practical capabilities of LLMs

Using large teacher models to guide the training of smaller student models has become the prevailing paradigm for efficient and effective learning. However, vocabulary mismatches between teacher and student language models pose significant challenges in language modeling, resulting in divergent token sequences and output distributions. To overcome these limitations, we propose Vocabulary-agnostic Teacher Guided Language Modeling (VocAgnoLM), a novel approach that bridges the gap caused by vocabulary mismatch through two key methods: (1) Token-level Lexical Alignment, which aligns token sequences across mismatched vocabularies, and (2) Teacher Guided Loss, which leverages the loss of teacher model to guide effective student training. We demonstrate its effectiveness in language modeling with 1B student model using various 7B teacher models with different vocabularies. Notably, with Qwen2.5-Math-Instruct, a teacher model sharing only about 6% of its vocabulary with TinyLlama, VocAgnoLM achieves a 46% performance improvement compared to naive continual pretraining. Furthermore, we demonstrate that VocAgnoLM consistently benefits from stronger teacher models, providing a robust solution to vocabulary mismatches in language modeling.

Haebin Shin, Lei Ji, Yeyun Gong, Sungdong Kim, Eunbi Choi, Minjoon Seo. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Inspired by the impressive reasoning capabilities demonstrated by reinforcement learning approaches like DeepSeek-R1, recent emerging research has begun exploring the use of reinforcement learning (RL) to enhance vision-language models (VLMs) for multimodal reasoning tasks. However, most existing multimodal reinforcement learning approaches remain limited to spatial reasoning within single-image contexts, yet still struggle to generalize to more complex and real-world scenarios involving multi-image positional reasoning, where understanding the relationships across images is crucial. To address this challenge, we propose a general reinforcement learning approach PeRL tailored for interleaved multimodal tasks, and a multi-stage strategy designed to enhance the exploration-exploitation trade-off, thereby improving learning efficiency and task performance. Specifically, we introduce permutation of image sequences to simulate varied positional relationships to explore more spatial and positional diversity. Furthermore, we design a rollout filtering mechanism for resampling to focus on trajectories that contribute most to learning optimal behaviors to exploit learned policies effectively. We evaluate our model on 5 widely-used multi-image benchmarks and 3 single-image benchmarks. Our experiments confirm that PeRL trained model consistently surpasses R1-related and interleaved VLM baselines by a large margin, achieving state-of-the-art performance on multi-image benchmarks, while preserving comparable performance on single-image tasks.

Panorama synthesis endeavors to craft captivating 360-degree visual landscapes, immersing users in the heart of virtual worlds. Nevertheless, contemporary panoramic synthesis techniques grapple with the challenge of semantically guiding the content generation process. Although recent breakthroughs in visual synthesis have unlocked the potential for semantic control in 2D flat images, a direct application of these methods to panorama synthesis yields distorted content. In this study, we unveil an innovative framework for generating high-resolution panoramas, adeptly addressing the issues of spherical distortion and edge discontinuity through sophisticated spherical modeling. Our pioneering approach empowers users with semantic control, harnessing both image and text inputs, while concurrently streamlining the generation of high-resolution panoramas using parallel decoding. We rigorously evaluate our methodology on a diverse array of indoor and outdoor datasets, establishing its superiority over recent related work, in terms of both quantitative and qualitative performance metrics. Our research elevates the controllability, efficiency, and fidelity of panorama synthesis to new levels.

Graphical User Interface (GUI) automation holds significant promise for assisting users with complex tasks, thereby boosting human productivity. Existing works leveraging Large Language Model (LLM) or LLM-based AI agents have shown capabilities in automating tasks on Android and Web platforms. However, these tasks are primarily aimed at simple device usage and entertainment operations. This paper presents a novel benchmark, Assistgui, to evaluate whether models are capable of manipulating the mouse and keyboard on the Windows platform in response to user-requested tasks. We carefully collected a set of 100 tasks from nine widely-used software applications, such as, After Effects and MS Word, each accompanied by the necessary project files for better evaluation. Moreover, we propose a multi-agent collaboration framework, which incorporates four agents to perform task decomposition, GUI parsing, action generation, and reflection. Our experimental results reveal that our multi-agent collaboration mechanism outshines existing methods in performance. Nevertheless, the potential remains substantial, with the best model attaining only a 46% success rate on our benchmark. We conclude with a thorough analysis of the current methods' limitations, setting the stage for future breakthroughs in this domain.

Representation learning is all about discovering the hidden modular attributes that generate the data faithfully. We explore the potential of Denoising Diffusion Probabilistic Model (DM) in unsupervised learning of the modular attributes. We build a theoretical framework that connects the diffusion time-steps and the hidden attributes, which serves as an effective inductive bias for unsupervised learning. Specifically, the forward diffusion process incrementally adds Gaussian noise to samples at each time-step, which essentially collapses different samples into similar ones by losing attributes, e.g., fine-grained attributes such as texture are lost with less noise added (i.e., early time-steps), while coarse-grained ones such as shape are lost by adding more noise (i.e., late time-steps). To disentangle the modular attributes, at each time-step t, we learn a t-specific feature to compensate for the newly lost attribute, and the set of all {1,...,t}-specific features, corresponding to the cumulative set of lost attributes, are trained to make up for the reconstruction error of a pre-trained DM at time-step t. On CelebA, FFHQ, and Bedroom datasets, the learned feature significantly improves attribute classification and enables faithful counterfactual generation, e.g., interpolating only one specified attribute between two images, validating the disentanglement quality. Codes are in https://github.com/yue-zhongqi/diti.

In recent years, the integration of vision and language understanding has led to significant advancements in artificial intelligence, particularly through Vision-Language Models (VLMs). However, existing VLMs face challenges in handling real-world applications with complex scenes and multiple objects, as well as aligning their focus with the diverse attention patterns of human users. In this paper, we introduce gaze information, feasibly collected by ubiquitous wearable devices such as MR glasses, as a proxy for human attention to guide VLMs. We propose a novel approach, Voila-A, for gaze alignment to enhance the effectiveness of these models in real-world applications. First, we collect hundreds of minutes of gaze data to demonstrate that we can mimic human gaze modalities using localized narratives. We then design an automatic data annotation pipeline utilizing GPT-4 to generate the VOILA-COCO dataset. Additionally, we introduce a new model VOILA-A that integrate gaze information into VLMs while maintain pretrained knowledge from webscale dataset. We evaluate Voila-A using a hold-out validation set and a newly collected VOILA-GAZE testset, which features real-life scenarios captured with a gaze-tracking device. Our experimental results demonstrate that Voila-A significantly outperforms several baseline models. By aligning model attention with human gaze patterns, Voila-A paves the way for more intuitive, user-centric VLMs and fosters engaging human-AI interaction across a wide range of applications.

Zhijian Hou, Wanjun Zhong, Lei Ji, Difei Gao, Kun Yan, W.k. Chan, Chong-Wah Ngo, Mike Zheng Shou, Nan Duan. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

To build Video Question Answering (VideoQA) systems capable of assisting humans in daily activities, seeking answers from long-form videos with diverse and complex events is a must. Existing multi-modal VQA models achieve promising performance on images or short video clips, especially with the recent success of large-scale multi-modal pre-training. However, when extending these methods to long-form videos, new challenges arise. On the one hand, using a dense video sampling strategy is computationally prohibitive. On the other hand, methods relying on sparse sampling struggle in scenarios where multi-event and multi-granularity visual reasoning are required. In this work, we introduce a new model named$\mathcal{M}ulti{-}$· modal Iterative$\mathcal{S}$.patial-temporal Transformer$(\mathcal{MIST})$) to better adapt pre-trained models for long-form VideoQA. Specifically,$\mathcal{MIST}$decomposes traditional dense spatial-temporal self-attention into cascaded segment and region selection modules that adaptively select frames and image regions that are closely relevant to the question itself. Visual concepts at different granularities are then processed efficiently through an attention module. In addition,$\mathcal{MIST}$iteratively conducts selection and attention over multiple layers to support reasoning over multiple events. The experimental results on four VideoQA datasets, including AGQA, NExT-QA, STAR, and Env-QA, show that$\mathcal{MIST}$achieves state-of-the-art performance and is superior at efficiency. The code is available at github.com/showlab/mist.

Electronic Health Records (EHRs), which contain patients' medical histories in various multi-modal formats, often overlook the potential for joint reasoning across imaging and table modalities underexplored in current EHR Question Answering (QA) systems. In this paper, we introduce EHRXQA, a novel multi-modal question answering dataset combining structured EHRs and chest X-ray images. To develop our dataset, we first construct two uni-modal resources: 1) The MIMIC- CXR-VQA dataset, our newly created medical visual question answering (VQA) benchmark, specifically designed to augment the imaging modality in EHR QA, and 2) EHRSQL (MIMIC-IV), a refashioned version of a previously established table-based EHR QA dataset. By integrating these two uni-modal resources, we successfully construct a multi-modal EHR QA dataset that necessitates both uni-modal and cross-modal reasoning. To address the unique challenges of multi-modal questions within EHRs, we propose a NeuralSQL-based strategy equipped with an external VQA API. This pioneering endeavor enhances engagement with multi-modal EHR sources and we believe that our dataset can catalyze advances in real-world medical scenarios such as clinical decision-making and research. EHRXQA is available at https://github.com/baeseongsu/ehrxqa.

Temporal Video Segmentation (TVS) is a fundamental video understanding task and has been widely researched in recent years. There are two subtasks of TVS: Video Action Segmentation (VAS) and Video Procedure Segmentation (VPS): VAS aims to recognize what actions happen in-side the video while VPS aims to segment the video into a sequence of video clips as a procedure. The VAS task inevitably relies on pre-defined action labels and is thus hard to scale to various open-domain videos. To overcome this limitation, the VPS task tries to divide a video into several category-independent procedure segments. However, the existing dataset for the VPS task is small (2k videos) and lacks diversity (only cooking domain). To tackle these problems, we collect a large and diverse dataset called TIPS, specifically for the VPS task. TIPS contains 63k videos including more than 300k procedure segments from instructional videos on YouTube, which covers plenty of how-to areas such as cooking, health, beauty, parenting, gardening, etc. We then propose a multi-modal Transformer with Gaussian Boundary Detection (MT-GBD) model for VPS, with the backbone of the Transformer and Convolution. Furthermore, we propose a new EIOU metric for the VPS task, which helps better evaluate VPS quality in a more comprehensive way. Experimental results show the effectiveness of our proposed model and metric.

Lei Ji, Xianglin Guo, Haoyang Huang, Xilin Chen. 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.

Kun Yan, Lei Ji, Huaishao Luo, Ming Zhou, Nan Duan, Shuai Ma. 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.

The task of video-based commonsense captioning aims to generate event-wise captions and meanwhile provide multiple commonsense descriptions (e.g., attribute, effect and intention) about the underlying event in the video. Prior works explore the commonsense captions by using separate networks for different commonsense types, which is time-consuming and lacks mining the interaction of different commonsense. In this paper, we propose a Hybrid Reasoning Network (HybridNet) to endow the neural networks with the capability of semantic-level reasoning and word-level reasoning. Firstly, we develop multi-commonsense learning for semantic-level reasoning by jointly training different commonsense types in a unified network, which encourages the interaction between the clues of multiple commonsense descriptions, event-wise captions and videos. Then, there are two steps to achieve the word-level reasoning: (1) a memory module records the history predicted sequence from the previous generation processes; (2) a memory-routed multi-head attention (MMHA) module updates the word-level attention maps by incorporating the history information from the memory module into the transformer decoder for word-level reasoning. Moreover, the multimodal features are used to make full use of diverse knowledge for commonsense reasoning. Experiments and abundant analysis on the large-scale Video-to-Commonsense benchmark show that our HybridNet achieves state-of-the-art performance compared with other methods.

Recent advances in the video question answering (i.e., VideoQA) task have achieved strong success by following the paradigm of fine-tuning each clip-text pair independently on the pretrained transformer-based model via supervised learning. Intuitively, multiple samples (i.e., clips) should be interdependent to capture similar visual and key semantic information in the same video. To consider the interdependent knowledge between contextual clips into the network inference, we propose a Siamese Sampling and Reasoning (SiaSamRea) approach, which consists of a siamese sampling mechanism to generate sparse and similar clips (i.e., siamese clips) from the same video, and a novel reasoning strategy for integrating the interdependent knowledge between contextual clips into the network. The reasoning strategy contains two modules: (1) siamese knowledge generation to learn the inter-relationship among clips; (2) siamese knowledge reasoning to produce the refined soft label by propagating the weights of inter-relationship to the predicted candidates of all clips. Finally, our SiaSamRea can endow the current multimodal reasoning paradigm with the ability of learning from inside via the guidance of soft labels. Extensive experiments demonstrate our SiaSamRea achieves state-of-the-art performance on five VideoQA benchmarks, e.g., a significant +2.1% gain on MSRVTT-QA, +2.9% on MSVD-QA, +1.0% on ActivityNet-QA, +1.8% on How2QA and +4.3% (action) on TGIF-QA.

Video captioning is a fundamental task for visual understanding. Previous works employ end-to-end networks to learn from the low-level vision feature and generate descriptive captions, which are hard to recognize fine-grained objects and lacks the understanding of crucial semantic concepts. According to DPC [19], these concepts generally present in the narrative transcripts of the instructional videos. The incorporation of transcript and video can improve the captioning performance. However, DPC directly concatenates the embedding of transcript with video features, which is incapable of fusing language and vision features effectively and leads to the temporal mis-alignment between transcript and video. This motivates us to 1) learn the semantic concepts explicitly and 2) design a temporal alignment mechanism to better align the video and transcript for the captioning task. In this paper, we start with an encoder-decoder backbone using transformer models. Firstly, we design a semantic concept prediction module as a multi-task to train the encoder in a supervised way. Then, we develop an attention based cross-modality temporal alignment method that combines the sequential video frames and transcript sentences. Finally, we adopt a copy mechanism to enable the decoder(generation) module to copy important concepts from source transcript directly. The extensive experimental results demonstrate the effectiveness of our model, which achieves state-of-the-art results on YouCookII dataset.

Understanding narrated instructional videos is important for both research and real-world web applications.Motivated by video dense captioning, we propose a model to generate procedure captions from narrated instructional videos which are a sequence of stepwise clips with description.Previous works on video dense captioning learn video segments and generate captions without considering transcripts.We argue that transcripts in narrated instructional videos can enhance video representation by providing fine-grained complimentary and semantic textual information.In this paper, we introduce a framework to ( 1) extract procedures by a cross-modality module, which fuses video content with the entire transcript; and (2) generate captions by encoding video frames as well as a snippet of transcripts within each extracted procedure.Experiments show that our model can achieve state-of-the-art performance in procedure extraction and captioning, and the ablation studies demonstrate that both the video frames and the transcripts are important for the task.

Image-text matching is a vital cross-modality task in artificial intelligence and has attracted increasing attention in recent years. Existing works have shown that learning semantic concepts is useful to enhance image representation and can significantly improve the performance of both image-to-text and text-to-image retrieval. However, existing models simply detect semantic concepts from a given image, which are less likely to deal with long-tail and occlusion concepts. Frequently co-occurred concepts in the same scene, e.g. bedroom and bed, can provide common-sense knowledge to discover other semantic-related concepts. In this paper, we develop a Scene Concept Graph (SCG) by aggregating image scene graphs and extracting frequently co-occurred concept pairs as scene common-sense knowledge. Moreover, we propose a novel model to incorporate this knowledge to improve image-text matching. Specifically, semantic concepts are detected from images and then expanded by the SCG. After learning to select relevant contextual concepts, we fuse their representations with the image embedding feature to feed into the matching module. Extensive experiments are conducted on Flickr30K and MSCOCO datasets, and prove that our model achieves state-of-the-art results due to the effectiveness of incorporating the external SCG.

Recently, Visual Question Answering (VQA) has emerged as one of the most significant tasks in multimodal learning as it requires understanding both visual and textual modalities. Existing methods mainly rely on extracting image and question features to learn their joint feature embedding via multimodal fusion or attention mechanism. Some recent studies utilize external VQA-independent models to detect candidate entities or attributes in images, which serve as semantic knowledge complementary to the VQA task. However, these candidate entities or attributes might be unrelated to the VQA task and have limited semantic capacities. To better utilize semantic knowledge in images, we propose a novel framework to learn visual relation facts for VQA. Specifically, we build up a Relation-VQA (R-VQA) dataset based on the Visual Genome dataset via a semantic similarity module, in which each data consists of an image, a corresponding question, a correct answer and a supporting relation fact. A well-defined relation detector is then adopted to predict visual question-related relation facts. We further propose a multi-step attention model composed of visual attention and semantic attention sequentially to extract related visual knowledge and semantic knowledge. We conduct comprehensive experiments on the two benchmark datasets, demonstrating that our model achieves state-of-the-art performance and verifying the benefit of considering visual relation facts.

Question retrieval, which aims to find similar versions of a given question, is playing a pivotal role in various question answering (QA) systems. This task is quite challenging, mainly in regard to five aspects: synonymy, polysemy, word order, question length, and data sparsity. In this article, we propose a unified framework to simultaneously handle these five problems. We use the word combined with corresponding concept information to handle the synonymy problem and the polysemous problem. Concept embedding and word embedding are learned at the same time from both the context-dependent and context-independent views. To handle the word-order problem, we propose a high-level feature-embedded convolutional semantic model to learn question embedding by inputting concept embedding and word embedding. Due to the fact that the lengths of some questions are long, we propose a value-based convolutional attentional method to enhance the proposed high-level feature-embedded convolutional semantic model in learning the key parts of the question and the answer. The proposed high-level feature-embedded convolutional semantic model nicely represents the hierarchical structures of word information and concept information in sentences with their layer-by-layer convolution and pooling. Finally, to resolve data sparsity, we propose using the multi-view learning method to train the attention-based convolutional semantic model on question–answer pairs. To the best of our knowledge, we are the first to propose simultaneously handling the above five problems in question retrieval using one framework. Experiments on three real question-answering datasets show that the proposed framework significantly outperforms the state-of-the-art solutions.

Deep Generative Models (DGMs) are able to extract high-level representations from massive unlabeled data and are explainable from a probabilistic perspective. Such characteristics favor sequence modeling tasks. However, it still remains a huge challenge to model sequences with DGMs. Unlike real-valued data that can be directly fed into models, sequence data consist of discrete elements and require being transformed into certain representations first. This leads to the following two challenges. First, high-level features are sensitive to small variations of inputs as well as the way of representing data. Second, the models are more likely to lose long-term information during multiple transformations. In this paper, we propose a Hierarchical Deep Generative Model With Dual Memory to address the two challenges. Furthermore, we provide a method to efficiently perform inference and learning on the model. The proposed model extends basic DGMs with an improved hierarchically organized multi-layer architecture. Besides, our model incorporates memories along dual directions, respectively denoted as broad memory and deep memory. The model is trained end-to-end by optimizing a variational lower bound on data log-likelihood using the improved stochastic variational method. We perform experiments on several tasks with various datasets and obtain excellent results. The results of language modeling show our method significantly outperforms state-of-the-art results in terms of generative performance. Extended experiments including document modeling and sentiment analysis, prove the high-effectiveness of dual memory mechanism and latent representations. Text random generation provides a straightforward perception for advantages of our model.

The question retrieval, which aims to find similar questions of a given question, is playing pivotal role in various question answering (QA) systems. This task is quite challenging mainly on three aspects: lexical gap, polysemy and word order. In this paper, we propose a unified framework to simultaneously handle these three problems. We use word combined with corresponding concept information to handle the polysemous problem. The concept embedding and word embedding are learned at the same time from both context-dependent and context-independent view. The lexical gap problem is handled since the semantic information has been encoded into the embedding. Then, we propose to use a high-level feature embedded convolutional semantic model to learn the question embedding by inputting the concept embedding and word embedding without manually labeling training data. The proposed framework nicely represent the hierarchical structures of word information and concept information in sentences with their layer-by-layer composition and pooling. Finally, the framework is trained in a weakly-supervised manner on question answer pairs, which can be directly obtained without manually labeling. Experiments on two real question answering datasets show that the proposed framework can significantly outperform the state-of-the-art solutions.

Knowledge based question answering (KBQA) has attracted much attention from both academia and industry in the field of Artificial Intelligence. However, many existing knowledge bases (KBs) are built by static triples. It is hard to answer user questions with different conditions, which will lead to significant answer variances in questions with similar intent. In this work, we propose to extract conditional knowledge base (CKB) from user question-answer pairs for answering user questions with different conditions through dialogue. Given a subject, we first learn user question patterns and conditions. Then we propose an embedding based co-clustering algorithm to simultaneously group the patterns and conditions by leveraging the answers as supervisor information. After that, we extract the answers to questions conditioned on both question pattern clusters and condition clusters as a CKB. As a result, when users ask a question without clearly specifying the conditions, we use dialogues in natural language to chat with users for question specification and answer retrieval. Experiments on real question answering (QA) data show that the dialogue model using automatically extracted CKB can more accurately answer user questions and significantly improve user satisfaction for questions with missing conditions.

According to the website AcronymFinder.com which is one of the world's largest and most comprehensive dictionaries of acronyms, an average of 37 new human-edited acronym definitions are added every day. There are 379,918 acronyms with 4,766,899 definitions on that site up to now, and each acronym has 12.5 definitions on average. It is a very important research topic to identify what exactly an acronym means in a given context for document comprehension as well as for document retrieval. In this paper, we propose two word embedding based models for acronym disambiguation. Word embedding is to represent words in a continuous and multidimensional vector space, so that it is easy to calculate the semantic similarity between words by calculating the vector distance. We evaluate the models on MSH Dataset and ScienceWISE Dataset, and both models outperform the state-of-art methods on accuracy. The experimental results show that word embedding helps to improve acronym disambiguation.

A frequent behavior of internet users is to compare among various comparable entities for decision making. As an instance, a user may compare among iPhone 5, Lumia 920 etc. products before deciding which cellphone to buy. However, it is a challenging problem to know what entities are generally comparable from the users' viewpoints in the open domain Web. In this paper, we propose a novel solution, which is known as Comparable Entity Graph Mining (CEGM), to learn an open-domain comparable entity graph from the user search queries. CEGM firstly mine seed comparable entity pairs from user search queries automatically using predefined query patterns. Next, it discovers more entity pairs with a confidence classifier in a bootstrapping fashion. Newly discovered entity pairs are organized into an open-domain comparable entity graph. Based on our empirical study over 1 billion queries of a commercial search engine, we build a comparable entity graph which covers 73.4% queries in the top 50 million unique queries of a commercial search engine. Through manual labeling in sampled sub-graphs, the average precision of comparable entities is 89.4%. As applications of the learned entity graph, the entity recommendation in Web search is empirically studied.

Advanced e-applications require comprehensive knowledge about their users’ preferences in order to provide accurate personalized services. In this paper, we propose to learn users’ preferences to product brands from their implicit feedbacks such as their searching and browsing behaviors in user Web browsing log data. The user brand preference learning problem is challenge since (1) the users’ implicit feedbacks are extremely sparse in various product domains; and (2) we can only observe positive feedbacks from users’ behaviors. In this paper, we propose a latent factor model to collaboratively mine users’ brand preferences across multiple domains simultaneously. By collective learning, the learning processes in all the domains are mutually enhanced and hence the problem of data scarcity in each single domain can be effectively addressed. On the other hand, we learn our model with an adaption of the Bayesian personalized ranking (BPR) optimization criterion which is a general learning framework for collaborative filtering from implicit feedbacks. Experiments with both synthetic and real world datasets show that our proposed model significantly outperforms the baselines.

Search engine users often have clear search tasks hidden behind their queries. Inspired by this, the modern search engines are providing an increasing number of services to help users simplify their key tasks. However, the problem of what are the major user search tasks with high traffic for which search engines should design special services is still underexplored. In this paper, we propose a novel Multi-view Random Walk (MRW) algorithm to measure the search task oriented similarity between queries, and then group search queries with similar tasks so that the major search tasks of users can be identified from search engine click-through log. The proposed MRW, which is a general framework to combine knowledge from different views in a random walk process, allows the random surfer to walk across different views to integrate information for search task discovery. Experimental results on click-through log of a commonly used commercial search engine show that our proposed MRW algorithm can effectively discover user search tasks.

Simplifying the key tasks of search engine users by directly retrieving to them structured knowledge according to their queries is attracting much attention from both industry and academia. A bottleneck of this challenging problem is how to extract the structured knowledge from the noisy and complex Web scale websites automatically. In this paper, we propose an unsupervised automatic wrapper induction algorithm, named as Scalable Knowledge Extractor from webSites (SKES). SKES induces the wrapper in a divide and conquer mode, i.e., it divides the general wrapper into several sub-wrappers to learn from the data independently. Moreover, through employing techniques such as tag path representation of Web pages, SKES is verified to be efficient and noise-tolerant by the experimental results. Furthermore, based on our automatically extracted knowledge, we also built a prototype to serve structured knowledge to end users for simplifying their key search tasks. Very positive feedbacks were received on the prototype.

Understanding search intents of users through their condensed short queries has attracted much attention both in academia and industry. The search intents of users are generally assumed to be associated with various query patterns, such as "MobileName price", where "MobileName" could be any named entity of mobile phone model and this pattern indicates that the user intends to buy a mobile phone. However, discovering the query intent patterns for general search is challenging mainly due to the difficulty in collecting sufficient training data for learning query patterns across a large number of searchable domains. In this work, we propose Cross Domain Random Walk (CDRW) algorithm, which is semi-supervised, to discover the query intent patterns across different domains from search engine click-through log data. Starting with some manually tagged seed queries in one or more independent domains, CDRW takes the query patterns as bridge and propagates the transition probability across domains to collect the query intent patterns among different domains based on the assumption that "users who have similar intent in different but similar domains will have high probability to share similar query patterns across domains". Different from classical random walk algorithms, CDRW walks across different domains to disseminate the shared knowledge in a transfer learning manner. Extensive experiment results on real log data of a commercial search engine well validate the effectiveness and efficiency of the proposed algorithm.

Understanding user intent from her sequential search behaviors, i.e. predicting the intent of each user query in a search session, is crucial for modern Web search engines. However, due to the huge number of user behavior variables and coarse level intent labels defined by human editors, it is very difficult to directly model user behavioral dynamics or user intent dynamics in user search sessions. In this paper, we propose a novel Sparse Hidden-Dynamic Conditional Random Fields (SHDCRF) model for user intent learning from their search sessions. Through incorporating the proposed hidden state variables, SHDCRF aims to learn a substructure, i.e. a set of related hidden variables, for each intent label and they are used to model the intermediate dynamics between user intent labels and user behavioral variables. In addition, SHDCRF learns a sparse relation between the hidden variables and intent labels to make the hidden state variables explainable. Extensive experiment results, on real user search sessions from a popular commercial search engine show that the proposed SHDCRF model significantly outperforms in terms of intent prediction results that those classical solutions such as Support Vector Machine (SVM), Conditional Random Field (CRF) and Latnet-Dynamic Conditional Random Field (LDCRF).

E-Commerce has shown its exponentially-growing business value in the past decade. However, in contrast to the successful examples in online sales, such as Amazon1 and eBay2, the online barter business is still underexplored due to the lack of corresponding information aggregation service. In this paper, we design and implement a novel vertical search engine, called ExSearch, to aggregate online barter information for developing the barter market. Different from classical general purpose Web search engines, ExSearch adopts a focused crawler to gather related information from various websites. We propose to automatically extract the barter information from free-text Web pages such that the unstructured information is represented in structured databases. In addition, we utilize the data mining techniques such as regression to fulfill the missing information, which cannot be extracted from the Web pages. Finally, we validate and rank the search results according to user queries. Experimental results show that each component module in our proposed ExSearch system is efficient and effective. The volunteer users are satisfied by and interested in this novel vertical search engine.

Both search engine click-through log and social annotation have been utilized as user feedback for search result re-ranking. However, to our best knowledge, no previous study has explored the correlation between these two factors for the task of search result ranking. In this paper, we show that the gap between search queries and social tags of the same Web page can well reflect its user preference score. Motivated by this observation, we propose a novel algorithm, called Query-Tag-Gap (QTG), to re-rank search results for better user satisfaction. Intuitively, on one hand, the search users' intentions are generally described by their queries before they read the search results. On the other hand, the Web annotators semantically tag Web pages after they read the content of the pages. The difference between users' recognition of the same page before and after they read it is a good reflection of user satisfaction. In this extended abstract, we formally define the query set and tag set of the same page as users' pre- and post- knowledge respectively. We empirically show the strong correlation between user satisfaction and user's knowledge gap before and after reading the page. Based on this gap, experiments have shown outstanding performance of our proposed QTG algorithm in search result re-ranking.

In recent years, different commercial Weblog subscribing systems have been proposed to return stories from users. subscribed feeds. In this paper, we propose a novel clustering-based RSS aggregator called as RSS Clusgator System (RCS) for Weblog reading. Note that an RSS feed may have several different topics. A user may only be interested in a subset of these topics. In addition there could be many different stories from multiple RSS feeds, which discuss similar topic from different perspectives. A user may be interested in this topic but do not know how to collect all feeds related to this topic. In contrast to many previous works, we cluster all stories in RSS feeds into hierarchical structure to better serve the readers. Through this way, users can easily find all their interested stories. To make the system current, we propose a flexible time window for incremental clustering. RCS utilizes both link information and content information for efficient clustering. Experiments show the effectiveness of RCS.

In this paper, we propose a novel ranking scheme named Affinity Ranking (AR) to re-rank search results by optimizing two metrics: (1) diversity -- which indicates the variance of topics in a group of documents; (2) information richness -- which measures the coverage of a single document to its topic. Both of the two metrics are calculated from a directed link graph named Affinity Graph (AG). AG models the structure of a group of documents based on the asymmetric content similarities between each pair of documents. Experimental results in Yahoo! Directory, ODP Data, and Newsgroup data demonstrate that our proposed ranking algorithm significantly improves the search performance. Specifically, the algorithm achieves 31% improvement in diversity and 12% improvement in information richness relatively within the top 10 search results.