Recent advances in long-context language models (LCLMs), designed to handle extremely long contexts, primarily focus on utilizing external contextual information, often leaving the influence of language models' parametric knowledge underexplored. In this work, we firstly investigate how this parametric knowledge affects content generation and demonstrate that its impact becomes increasingly pronounced as context length extends. Furthermore, we show that the model’s ability to utilize parametric knowledge, which we call parametric recall ability, does not improve simultaneously with its ability to leverage contextual knowledge through extrinsic retrieval ability. Moreover, better extrinsic retrieval ability can interfere with the model’s parametric recall ability, limiting its full potential. To bridge this gap, we design a simple yet effective Hybrid Needle-in-a-Haystack test that evaluates models based on their capabilities across both abilities, rather than solely emphasizing extrinsic retrieval ability. Our experimental results reveal that Qwen-2.5 models significantly outperform Llama-3.1 models, demonstrating superior potential to combine various abilities. Moreover, even the more powerful Llama-3.1-70B-Instruct model fails to exhibit better performance, highlighting the importance of evaluating models from a dual-ability perspective.

Yuxuan Lu, Jing Huang, Yan Han, Bingsheng Yao, Sisong Bei, Yaochen Xie, Yisi Sang, Qi He, Dakuo Wang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Spatio-temporal reasoning in time series involves the explicit synthesis of temporal dynamics, spatial dependencies, and textual context.This capability is vital for high-stakes decision-making in systems such as traffic networks, power grids, and disease propagation.However, the field remains underdeveloped because most existing works prioritize predictive accuracy over reasoning.To address the gap, we introduce ST-Bench, a benchmark consisting of four core tasks, including etiological reasoning, entity identification, correlation reasoning, and in-context forecasting, developed via a network SDE-based multi-agent data synthesis pipeline.We then propose STReasoner, which empowers LLM to integrate time series, graph structure, and text for explicit reasoning.To promote spatially grounded logic, we introduce S-GRPO, a reinforcement learning algorithm that rewards performance gains specifically attributable to spatial information.Experiments show that STReasoner achieves average accuracy gains between 17% and 135% at only 0.004× the cost of proprietary models and generalizes robustly to real-world data.Our code is available at https://github.com/LingFengGold/STReasoner.

Weiqi Wang, Limeng Cui, Xin Liu, Sreyashi Nag, Wenju Xu, Chen Luo, Sheikh Muhammad Sarwar, Yang Li, Hansu Gu, Hui Liu, Changlong Yu, Jiaxin Bai, Yifan Gao, Haiyang Zhang, Qi He, Shuiwang Ji, Yangqiu Song. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Few-shot Chain-of-Thought (CoT) prompting has demonstrated strong performance in improving the reasoning capabilities of large language models (LLMs). While theoretical investigations have been conducted to understand CoT, the underlying transformer used in these studies isolates the CoT reasoning process into separated in-context learning steps (Stepwise ICL). In this work, we theoretically show that, compared to Stepwise ICL, the transformer gains better error correction ability and more accurate predictions if the reasoning from earlier steps (Coherent CoT) is integrated. Given that this coherent reasoning changes the behavior of the transformer, we further investigate the sensitivity of the transformer with Coherent CoT when the demonstration examples are corrupted at the inference stage. Our theoretical results indicate that the transformer is more sensitive to errors in intermediate reasoning steps than the final outcome. Building upon this observation, we propose an improvement on CoT by incorporating both correct and incorrect reasoning paths in the demonstration. Our experiments validate the effectiveness of the proposed approach.

Multimodal Large Language Models (MLLMs) trained on massive data may memorize sensitive personal information and photos, posing serious privacy risks.To mitigate this, MLLM unlearning methods are proposed, which finetune MLLMs to forget sensitive information.However, it remains unclear whether the knowledge has been truly forgotten or just hidden in the model.Therefore, we propose to study a novel problem of MLLM unlearning attack, which aims to recover the unlearned knowledge of an unlearned MLLM.To achieve the goal, we propose a novel framework-Stealthy Unlearning Attack (SUA)-that learns a universal noise pattern.When applied to input images, this noise can trigger the model to reveal unlearned content.While pixel-level perturbations may be visually subtle, they can be detected in the semantic embedding space, making such attacks vulnerable to potential defenses.To improve stealthiness, we introduce an embedding alignment loss that minimizes the difference between the perturbed and denoised image embeddings, ensuring that the attack remains semantically unnoticeable.Experimental results show that SUA can effectively recover unlearned information from MLLMs.Furthermore, the learned noise generalizes well-i.e., a single perturbation trained on a few samples can reveal forgotten contents in unseen images.

Large language models (LLMs) have shown remarkable proficiency in generating text, benefiting from extensive training on vast textual corpora. However, LLMs may also acquire unwanted behaviors from the diverse and sensitive nature of their training data, which can include copyrighted and private content. Machine unlearning has been introduced as a viable solution to remove the influence of such problematic content without the need for costly and time-consuming retraining. This process aims to erase specific knowledge from LLMs while preserving as much model utility as possible. Despite the effectiveness of current unlearning methods, little attention has been given to whether existing unlearning methods for LLMs truly achieve forgetting or merely hide the knowledge, which current unlearning benchmarks fail to detect. This paper reveals that applying quantization to models that have undergone unlearning can restore the "forgotten" information. We conduct comprehensive experiments using various quantization techniques across multiple precision levels to thoroughly evaluate this phenomenon. We find that for unlearning methods with utility constraints, the unlearned model retains an average of 21\% of the intended forgotten knowledge in full precision, which significantly increases to 83\% after 4-bit quantization. Based on our empirical findings, we provide a theoretical explanation for the observed phenomenon and propose a quantization-robust unlearning strategy aimed at mitigating this intricate issue. Our results highlight a fundamental tension between preserving the utility of the unlearned model and preventing knowledge recovery through quantization, emphasizing the challenge of balancing these two objectives. Altogether, our study underscores a major failure in existing unlearning methods for LLMs, strongly advocating for more comprehensive and robust strategies to ensure authentic unlearning without compromising model utility. Our code is available at: https://github.com/zzwjames/FailureLLMUnlearning.

Large Language Models (LLMs) are revolutionizing E-Commerce by enabling product recommendation, search, classification, question answering, and advertising applications. Their increasing adoption in real-world systems underscores their potential; however, challenges persist in ensuring accuracy, efficiency, fairness, and privacy. This workshop aims to bring together researchers and industry practitioners to explore both the limitations and opportunities of LLMs in e-commerce. The workshop seeks to foster collaboration, bridge the gap between academia and industry, and drive innovation in the application of LLMs to E-Commerce through discussions on model design, algorithmic advancements, and practical deployment.

Large language models (LLMs) based on Transformer architecture are powerful but face challenges with deployment, inference latency, and costly fine-tuning. These limitations highlight the emerging potential of small language models (SLMs), which can either replace LLMs through innovative architectures and technologies, or assist them as efficient proxy or reward models. Emerging architectures such as Mamba and xLSTM address the quadratic scaling of inference with window length in Transformers by enabling linear scaling. To maximize SLM performance, test-time compute scaling strategies reduce the performance gap with LLMs by allocating extra compute budget during test time. Beyond standalone usage, SLMs could also assist in LLMs via weak-to-strong learning, proxy tuning, and guarding, fostering secure and efficient LLM deployment. Lastly, the trustworthiness of SLMs remains a critical yet underexplored research area. However, there is a lack of tutorials on cutting-edge SLM technologies, prompting us to conduct one.

Ran Xu, Hui Liu, Sreyashi Nag, Zhenwei Dai, Yaochen Xie, Xianfeng Tang, Chen Luo, Yang Li, Joyce C. Ho, Carl Yang, Qi He. 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.

Although Large Language Models (LLMs) have demonstrated impressive capabilities across a wide range of tasks, growing concerns have emerged over the misuse of sensitive, copyrighted, or harmful data during training. To address these concerns, unlearning techniques have been developed to remove the influence of specific data without retraining from scratch. However, this paper reveals a critical vulnerability in fine-tuning-based unlearning: a malicious user can craft a manipulated forgetting request that stealthily degrades the model’s utility for benign users. We demonstrate this risk through a red-teaming Stealthy Attack (SA), which is inspired by two key limitations of existing unlearning—the inability to constrain the scope of unlearning effect and the failure to distinguish benign tokens from unlearning signals. Prior work has shown that unlearned models tend to memorize forgetting data as unlearning signals, and respond with hallucinations or feigned ignorance when unlearning signals appear in the input. By subtly increasing the presence of common benign tokens in the forgetting data, SA enhances the connection between benign tokens and unlearning signals. As a result, when normal users include such tokens in their prompts, the model exhibits unlearning behaviors, leading to unintended utility degradation. To address this vulnerability, we propose Scope-aware Unlearning (SU), a lightweight enhancement that introduces a scope term into the unlearning objective, encouraging the model to localize the forgetting effect. Our method requires no additional data processing, integrates seamlessly with existing fine-tuning frameworks, and significantly improves robustness against SA. Extensive experiments validate the effectiveness of both SA and SU.

Test-time scaling (TTS) enhances the performance of large language models (LLMs) by allocating additional compute resources during inference. However, existing research primarily investigates TTS in single-stage tasks; while many real-world problems are multi-stage complex tasks, composed of a sequence of heterogeneous subtasks with each subtask requires LLM of specific capability. Therefore, we study a novel problem: the test-time compute-optimal scaling in multi-stage complex tasks, aiming to select suitable models and allocate budgets per subtask to maximize overall performance. TTS in multi-stage tasks introduces two fundamental challenges: (i) The combinatorial search space of model and budget allocations, combined with the high cost of inference, makes brute-force search impractical. (ii) The optimal model and budget allocations across subtasks are interdependent, increasing the complexity of the compute-optimal search. To address this gap, we conduct extensive pilot experiments on four tasks across six datasets, deriving three empirical insights characterizing the behavior of LLMs in multi-stage complex tasks. Informed by these insights, we propose AgentTTS, an LLM-agent-based framework that autonomously searches for compute-optimal allocations through iterative feedback-driven interactions with the execution environment. Experimental results demonstrate that AgentTTS significantly outperforms traditional and other LLM-based baselines in search efficiency, and shows improved robustness to varying training set sizes and enhanced interpretability.

The rapid growth of online shopping stores, such as Amazon, has led to services reaching billions of people worldwide. With global retail sales exceeding $6 trillion in 2024, customer expectations for personalized and seamless shopping experiences have heightened. Traditional online shopping experiences, such as search and navigation systems, often fall short in addressing complex shopping journeys. Conversational shopping (such as Amazon Rufus) offers a transformative approach by enabling dynamic, multi-turn dialogues that closely resemble human interactions. This allows customers to explore product options, seek clarifications, and receive personalized recommendations, thereby enhancing product discovery and informed decision-making. In this paper, we share our year-long journey of using language models for conversational shopping at Amazon and introduce how we use LLM fine-tuning techniques to enhance LLMs for a conversational shopping experience like Amazon Rufus. We also introduce innovative strategies for training data collection and demonstrate real-world applications, including product recommendations, clarification mechanisms, and internationalization for global customers.

Prompting LLMs offers an efficient way to guide output generation without explicit model training. In the e-commerce domain, prompt-based applications are widely used in query understanding, recommender systems, and customer support. However, adapting LLMs to different tasks often requires extensive prompt engineering by domain experts, along with frequent updates to align with evolving business needs. Additionally, crafting fully unbiased natural language prompts remains a challenge for humans. To address these challenges, we propose a novel framework, Examples as the Prompt (EaP). Specifically, EaP automatically selects the most representative examples to maximize the few-shot capability of LLMs. It is efficient due to its unsupervised example selection and adaptive to potential data distribution shifts. We validate EaP on four real-world production use cases, demonstrating that it achieves comparable or even superior performance comparing to hand-crafted prompts designed by domain experts. Additionally, we introduce EaPlite, which entirely replaces the natural language components of prompts with labeled examples. EaPlite improves LLM inference speed by up to 70% without compromising performance. The online A/B test shows that using EaP and EaPlite for data labeling can bring significant composite revenue gain by 0.06%.

Large language models (LLMs) have demonstrated emergent abilities in text generation, question answering, and reasoning, facilitating various tasks and domains. Despite their proficiency in various tasks, LLMs like PaLM 540B and Llama-3.1 405B face limitations due to large parameter sizes and computational demands, often requiring cloud API use, which raises privacy concerns, limits real-time applications on edge devices, and increases fine-tuning costs. Additionally, LLMs often underperform in specialized domains such as healthcare and law due to insufficient domain-specific knowledge, necessitating specialized models. Therefore, Small Language Models (SLMs) are increasingly favored for their low inference latency, cost-effectiveness, efficient development, and easy customization and adaptability. These models are particularly well-suited for resource-limited environments and domain knowledge acquisition, addressing LLMs’ challenges and proving ideal for applications that require localized data handling for privacy, minimal inference latency for efficiency, and domain knowledge acquisition through lightweight fine-tuning. The rising demand for SLMs has spurred extensive research and development. However, a comprehensive survey investigating issues related to the definition, acquisition, application, enhancement, and reliability of SLM remains lacking, prompting us to conduct a detailed survey on these topics. The definition of SLMs varies widely; thus, to standardize, we propose defining SLMs by their capability to perform specialized tasks and suitability for resource-constrained settings, setting boundaries based on the minimal size for emergent abilities and the maximum size sustainable under resource constraints. For other aspects, we provide a taxonomy of relevant models/methods and develop general frameworks for each category to enhance and utilize SLMs effectively. We have compiled the collected SLM models and related methods on GitHub: https://github.com/FairyFali/SLMs-Survey .

Online shopping platforms, such as Amazon, offer services to billions of people worldwide. Unlike web search or other search engines, product search engines have their unique characteristics, primarily featuring short queries which are mostly a combination of product attributes and structured product search space. The uniqueness of product search underscores the crucial importance of the query understanding component. However, there are limited studies focusing on exploring this impact within real-world product search engines. In this work, we aim to bridge this gap by conducting a comprehensive study and sharing our year-long journey investigating how the query understanding service impacts Amazon Product Search. Firstly, we explore how query understanding-based ranking features influence the ranking process. Next, we delve into how the query understanding system contributes to understanding the performance of a ranking model. Building on the insights gained from our study on the evaluation of the query understanding-based ranking model, we propose a query understanding-based multi-task learning framework for ranking. We present our studies and investigations on Amazon Search.

Fairness in Graph Convolutional Neural Networks (GCNs) becomes a more and more important concern as GCNs are adopted in many crucial applications. Societal biases against sensitive groups may exist in many real world graphs. GCNs trained on those graphs may be vulnerable to being affected by such biases. In this paper, we adopt the well-known fairness notion of demographic parity and tackle the challenge of training fair and accurate GCNs efficiently. We present an in-depth analysis on how graph structure bias, node attribute bias, and model parameters may affect the demographic parity of GCNs. Our insights lead to FairSample, a framework that jointly mitigates the three types of biases. We employ two intuitive strategies to rectify graph structures. First, we inject edges across nodes that are in different sensitive groups but similar in node features. Second, to enhance model fairness and retain model quality, we develop a learnable neighbor sampling policy using reinforcement learning. To address the bias in node features and model parameters, FairSample is complemented by a regularization objective to optimize fairness.

The ability to infer an individual's expertise for a given skill has proven to be crucial in creating economic opportunity for every talent of the global workforce. Applications ranging from recommending relevant job opportunities to talents to providing better candidate suggestions to recruiters, all benefit from deep understanding of the skill "proficiency" of the talent pool.

The main challenge of adapting Graph convolutional networks (GCNs) to large-scale graphs is the scalability issue due to the uncontrollable neighborhood expansion in the aggregation stage. Several sampling algorithms have been proposed to limit the neighborhood expansion. However, these algorithms focus on minimizing the variance in sampling to approximate the original aggregation. This leads to two critical problems: 1) low accuracy because the sampling policy is agnostic to the performance of the target task, and 2) vulnerability to noise or adversarial attacks on the graph.

At LinkedIn, we want to create economic opportunity for everyone in the global workforce. To make this happen, LinkedIn offers a reactive Job Search system, and a proactive Jobs You May Be Interested In (JYMBII) system to match the best candidates with their dream jobs. One of the most challenging tasks for developing these systems is to properly extract important skill entities from job postings and then target members with matched attributes. In this work, we show that the commonly used text-based salience and market-agnostic skill extraction approach is sub-optimal because it only considers skill mention and ignores the salient level of a skill and its market dynamics, i.e., the market supply and demand influence on the importance of skills. To address the above drawbacks, we present Job2Skills, our deployed salience and market-aware skill extraction system. The proposed Job2Skills shows promising results in improving the online performance of job recommendation (JYMBII) (+1.92% job apply) and skill suggestions for job posters (-37% suggestion rejection rate). Lastly, we present case studies to show interesting insights that contrast traditional skill recognition method and the proposed Job2Skills from occupation, industry, country, and individual skill levels. Based on the above promising results, we deployed the Job2Skills online to extract job targeting skills for all 20M job postings served at LinkedIn.

As the world's largest professional network, LinkedIn wants to create economic opportunity for everyone in the global workforce. One of its most critical missions is matching jobs with processionals. Improving job targeting accuracy and hire efficiency align with LinkedIn's Member First Motto. To achieve those goals, we need to understand unstructured job postings with noisy information. We applied deep transfer learning to create domain-specific job understanding models. After this, jobs are represented by professional entities, including titles, skills, companies, and assessment questions. To continuously improve LinkedIn's job understanding ability, we designed an expert feedback loop where we integrated job understanding models into LinkedIn's products to collect job posters' feedback. In this demonstration, we present LinkedIn's job posting flow and demonstrate how the integrated deep job understanding work improves job posters' satisfaction and provides significant metric lifts in LinkedIn's job recommendation system.

At LinkedIn, we want to create economic opportunity for everyone in the global workforce. A critical aspect of this goal is matching jobs with qualified applicants. To improve hiring efficiency and reduce the need to manually screening each applicant, we develop a new product where recruiters can ask screening questions online so that they can filter qualified candidates easily. To add screening questions to all 20M active jobs at Linked In, we propose a new task that aims to automatically generate screening questions for a given job posting. To solve the task of generating screening questions, we develop a two-stage deep learning model called Job2Questions, where we apply a deep learning model to detect intent from the text description, and then rank the detected intents by their importance based on other contextual features. Since this is a new product with no historical data, we employ deep transfer learning to train complex models with limited training data. We launched the screening question product and our AI models to LinkedIn users and observed significant impact in the job marketplace. During our online A/B test, we observed +53.10% screening question suggestion acceptance rate, +22.17% job coverage, +190% recruiter-applicant interaction, and +11 Net Promoter Score. In sum, the deployed Job2Questions model helps recruiters to find qualified applicants and job seekers to find jobs they are qualified for.

Network representations have been shown to improve performance within a variety of tasks, including classification, clustering, and link prediction. However, most models either focus on moderate-sized, homogeneous networks or require a significant amount of auxiliary input to be provided by the user. Moreover, few works have studied network representations in real-world heterogeneous social networks with ambiguous social connections and are often incomplete. In the present work, we investigate the problem of learning low-dimensional node representations in heterogeneous professional social networks (HPSNs), which are incomplete and have ambiguous social connections. We present a general heterogeneous network representation learning model called Star2Vec that learns entity and person embeddings jointly using a social connection strength-aware biased random walk combined with a node-structure expansion function. Experiments on LinkedIn's Economic Graph and publicly available snapshots of Facebook's network show that Star2Vec outperforms existing methods on members' industry and social circle classification, skill and title clustering, and member-entity link predictions. We also conducted large-scale case studies to demonstrate practical applications of the Star2Vec embeddings trained on LinkedIn's Economic Graph such as next career move, alternative career suggestions, and general entity similarity searches.

The main mission of LinkedIn is to connect 610M+ members to the right opportunities. To find the right opportunities, LinkedIn needs to understand each member's skill set and their expertise levels accurately. However, estimating members' skill expertise is challenging due to lack of ground-truth. So far, the industry relied on either hand-created small scale data, or large scale social gestures containing a lot of social bias (e.g., endorsements).

LinkedIn dynamically delivers update activities from a user's interpersonal network to more than 300 million members in the personalized feed that ranks activities according their "relevance" to the user. This paper discloses the implementation details behind this personalized feed system at LinkedIn which can not be found from related work, and addresses the scalability and data sparsity challenges for deploying the system online. More specifically, we focus on the personalization models by generating three kinds of affinity scores: Viewer-ActivityType Affinity, Viewer-Actor Affinity, and Viewer-Actor-ActivityType Affinity. Extensive experiments based on online bucket tests (A/B experiments) and offline evaluation illustrate the effect of our personalization models in LinkedIn feed.

Twitter has attracted millions of users to share and disseminate most up-to-date information, resulting in large volumes of data produced everyday. However, many applications in Information Retrieval (IR) and Natural Language Processing (NLP) suffer severely from the noisy and short nature of tweets. In this paper, we propose a novel framework for tweet segmentation in a batch mode, called HybridSeg. By splitting tweets into meaningful segments, the semantic or context information is well preserved and easily extracted by the downstream applications. HybridSeg finds the optimal segmentation of a tweet by maximizing the sum of the stickiness scores of its candidate segments. The stickiness score considers the probability of a segment being a phrase in English (i.e., global context) and the probability of a segment being a phrase within the batch of tweets (i.e., local context). For the latter, we propose and evaluate two models to derive local context by considering the linguistic features and term-dependency in a batch of tweets, respectively. HybridSeg is also designed to iteratively learn from confident segments as pseudo feedback. Experiments on two tweet data sets show that tweet segmentation quality is significantly improved by learning both global and local contexts compared with using global context alone. Through analysis and comparison, we show that local linguistic features are more reliable for learning local context compared with term-dependency. As an application, we show that high accuracy is achieved in named entity recognition by applying segment-based part-of-speech (POS) tagging.

Graph has been a ubiquitous and essential data representation to model real world objects and their relationships. Today, large amounts of graph data have been generated by various applications. Graph summarization techniques are crucial in uncovering useful insights about the patterns hidden in the underlying data. However, all existing works in graph summarization are single-process solutions, and as a result cannot scale to large graphs. In this paper, we introduce three distributed graph summarization algorithms to address this problem. Experimental results show that the proposed algorithms can produce good quality summaries and scale well with increasing data sizes. To the best of our knowledge, this is the first work to study distributed graph summarization methods.

Users on an online social network site generate a large number of heterogeneous activities, ranging from connecting with other users, to sharing content, to updating their profiles. The set of activities within a user's network neighborhood forms a stream of updates for the user's consumption. In this paper, we report our experience with the problem of ranking activities in the LinkedIn homepage feed. In particular, we provide a taxonomy of social network activities, describe a system architecture (with a number of key components open-sourced) that supports fast iteration in model development, demonstrate a number of key factors for effective ranking, and report experimental results from extensive online bucket tests.

Understanding social tie development among users is crucial for user engagement in social networking services. In this paper, we analyze the social interactions, both online and offline, of users and investigate the development of their social ties using data trail of “how social ties grow” left in mobile and social networking services. To the best of our knowledge, this is the first research attempt on studying social tie development by considering both online and offline interactions in a heterogeneous yet realistic relationship. In this study, we aim to answer three key questions: 1) is there a correlation between online and offline interactions? 2) how is the social tie developed via heterogeneous interaction channels? 3) would the development of social tie between two users be affected by their common friends? To achieve our goal, we develop a Social-aware Hidden Markov Model (SaHMM) that explicitly takes into account the factor of common friends in measure of the social tie development. Our experiments show that, comparing with results obtained using HMM and other heuristic methods, the social tie development captured by our SaHMM is significantly more consistent to lifetime profiles of users.

introduction Free Access Share on Introduction to the Special Issue on Linking Social Granularity and Functions Authors: Qi He LinkedIn LinkedInView Profile , Juanzi Li Tsinghua University Tsinghua UniversityView Profile , Rong Yan Square SquareView Profile , John Yen Pennsylvania State University Pennsylvania State UniversityView Profile , Haizheng Zhang StarMerx LLC StarMerx LLCView Profile Authors Info & Claims ACM Transactions on Intelligent Systems and TechnologyVolume 5Issue 2Article No.: 22pp 1–3https://doi.org/10.1145/2594452Published:30 April 2014Publication History 2citation322DownloadsMetricsTotal Citations2Total Downloads322Last 12 Months4Last 6 weeks2 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Publisher SiteeReaderPDF

In this paper, we introduce “task trail” to understand user search behaviors. We define a task to be an atomic user information need, whereas a task trail represents all user activities within that particular task, such as query reformulations, URL clicks. Previously, web search logs have been studied mainly at session or query level where users may submit several queries within one task and handle several tasks within one session. Although previous studies have addressed the problem of task identification, little is known about the advantage of using task over session or query for search applications. In this paper, we conduct extensive analyses and comparisons to evaluate the effectiveness of task trails in several search applications: determining user satisfaction, predicting user search interests, and suggesting related queries. Experiments on large scale data sets of a commercial search engine show that: (1) Task trail performs better than session and query trails in determining user satisfaction; (2) Task trail increases webpage utilities of end users comparing to session and query trails; (3) Task trails are comparable to query trails but more sensitive than session trails in measuring different ranking functions; (4) Query terms from the same task are more topically consistent to each other than query terms from different tasks; (5) Query suggestion based on task trail is a good complement of query suggestions based on session trail and click-through bipartite. The findings in this paper verify the need of extracting task trails from web search logs and enhance applications in search and recommendation systems.

Twitter has attracted hundred millions of users to share and disseminate most up-to-date information. However, the noisy and short nature of tweets makes many applications in information retrieval (IR) and natural language processing (NLP) challenging. Recently, segment-based tweet representation has demonstrated effectiveness in named entity recognition (NER) and event detection from tweet streams. To split tweets into meaningful phrases or segments, the previous work is purely based on external knowledge bases, which ignores the rich local context information embedded in the tweets. In this paper, we propose a novel framework for tweet segmentation in a batch mode, called HybridSeg. HybridSeg incorporates local context knowledge with global knowledge bases for better tweet segmentation. HybridSeg consists of two steps: learning from off-the-shelf weak NERs and learning from pseudo feedback. In the first step, the existing NER tools are applied to a batch of tweets. The named entities recognized by these NERs are then employed to guide the tweet segmentation process. In the second step, HybridSeg adjusts the tweet segmentation results iteratively by exploiting all segments in the batch of tweets in a collective manner. Experiments on two tweet datasets show that HybridSeg significantly improves tweet segmentation quality compared with the state-of-the-art algorithm. We also conduct a case study by using tweet segments for the task of named entity recognition from tweets. The experimental results demonstrate that HybridSeg significantly benefits the downstream applications.

Even though queries received by traditional information retrieval systems are quite short, there are many application scenarios where long natural language queries are more effective. Further, incorporating term position information can help improve results of long queries. However, the techniques for incorporating term position information have been developed for terse queries and hence, can not be directly applied to long queries. Though there exist some methods for performing proximal search for long queries, they are not scalable due to long query response times. We describe an intuitive and simple, yet effective technique that implicitly incorporates term position information for long queries in a scalable manner. Our proposed approach achieves more than 700% faster query response times while maintaining the quality of retrieved results when compared with a state-of-the-art method for performing proximal search for very long queries.

Newly emerged event-based online social services, such as Meetup and Plancast, have experienced increased popularity and rapid growth. From these services, we observed a new type of social network - event-based social network (EBSN). An EBSN does not only contain online social interactions as in other conventional online social networks, but also includes valuable offline social interactions captured in offline activities. By analyzing real data collected from Meetup, we investigated EBSN properties and discovered many unique and interesting characteristics, such as heavy-tailed degree distributions and strong locality of social interactions.

Many private and/or public organizations have been reported to create and monitor targeted Twitter streams to collect and understand users' opinions about the organizations. Targeted Twitter stream is usually constructed by filtering tweets with user-defined selection criteria e.g. tweets published by users from a selected region, or tweets that match one or more predefined keywords. Targeted Twitter stream is then monitored to collect and understand users' opinions about the organizations. There is an emerging need for early crisis detection and response with such target stream. Such applications require a good named entity recognition (NER) system for Twitter, which is able to automatically discover emerging named entities that is potentially linked to the crisis. In this paper, we present a novel 2-step unsupervised NER system for targeted Twitter stream, called TwiNER. In the first step, it leverages on the global context obtained from Wikipedia and Web N-Gram corpus to partition tweets into valid segments (phrases) using a dynamic programming algorithm. Each such tweet segment is a candidate named entity. It is observed that the named entities in the targeted stream usually exhibit a gregarious property, due to the way the targeted stream is constructed. In the second step, TwiNER constructs a random walk model to exploit the gregarious property in the local context derived from the Twitter stream. The highly-ranked segments have a higher chance of being true named entities. We evaluated TwiNER on two sets of real-life tweets simulating two targeted streams. Evaluated using labeled ground truth, TwiNER achieves comparable performance as with conventional approaches in both streams. Various settings of TwiNER have also been examined to verify our global context + local context combo idea.

Link prediction is one of central tasks in the study of social network evolution and has many applications. In this paper, we use time series to describe node behavior, extract temporal features from the time series to characterize behavior evolution of nodes, and use the temporal features for link prediction. Our experimental results on several real datasets suggest that including the temporal features developed in the paper significantly improve link prediction performance.

Automatic recommendation of citations for a manuscript is highly valuable for scholarly activities since it can substantially improve the efficiency and quality of literature search. The prior techniques placed a considerable burden on users, who were required to provide a representative bibliography or to mark passages where citations are needed. In this paper we present a system that considerably reduces this burden: a user simply inputs a query manuscript (without a bibliography) and our system automatically finds locations where citations are needed. We show that naïve approaches do not work well due to massive noise in the document corpus. We produce a successful approach by carefully examining the relevance between segments in a query manuscript and the representative segments extracted from a document corpus. An extensive empirical evaluation using the CiteSeerX data set shows that our approach is effective.

Online social networking platforms have become a popular channel of communications among people. However, most people can only keep in touch with a limited number of friends. This phenomenon results in a low-connectivity social network in terms of communications, which is inefficient for information propagation and social engagement. In this paper, we introduce a new recommendation service, called link revival, that suggests users to re-connect with their old friends, such that the resulted connection will improve the social network connectivity. To achieve high connectivity improvement under the dynamic social network evolvement, we propose a graph prediction-based recommendation strategy, which selects proper candidates based on the prediction of their future behaviors. We then develop an effective model that exploits non-homogeneous Poisson process and second-order self-similarity in prediction. Through comprehensive experimental studies on two real datasets (Phone Call Network and Facebook Wall-posts), we demonstrate that our proposed approach can significantly increase the social network connectivity, and that the approach outperforms other baseline solutions. The results also show that our solution is more suitable for online social networks like Facebook, partially due to the stronger long range dependency and lower communication costs in the interactions.

Social networks mediate not only the relations between entities, but also the patterns of information propagation among them and their communication behavior. In this paper, we extensively study the temporal annotations (e.g., time stamps and duration) of historical communications in social networks and propose two novel tools -- communication motifs and maximum-flow communication motifs -- for characterizations of the patterns of information propagation in social networks. Using these motifs, we verify the following hypothesis in social communication network: 1) the functional behavioral patterns of information propagation within both social networks are stable over time; 2) the patterns of information propagation in synchronous and asynchronous social networks are different and sensitive to the cost of communication; and 3) the speed and the amount of information that is propagated through a network are correlated and dependent on individual profiles.

When micro logging becomes a very popular social media, finding interesting posts from high volume stream of user posts is a challenging research problem. To organize large number of posts, users can assign tags to posts so that these posts can be navigated and searched by tag. In this paper, we focus on modeling the interestingness of hash tags in Twitter, the largest and most active micro logging site. We propose to first construct communities based on both follow links and tagged interactions. We then measure the dispersion and divergence of users and tweets using hash tags among the constructed communities. The interestingness of hash tags are then derived from these community-based dispersion and divergence features. We further introduce a supervised approach to rank hash tags by interestingness. Our experiments on a Twitter dataset show that the proposed approach achieves a fairly good performance.

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.

When you write papers, how many times do you want to make some citations at a place but you are not sure which papers to cite? Do you wish to have a recommendation system which can recommend a small number of good candidates for every place that you want to make some citations? In this paper, we present our initiative of building a context-aware citation recommendation system. High quality citation recommendation is challenging: not only should the citations recommended be relevant to the paper under composition, but also should match the local contexts of the places citations are made. Moreover, it is far from trivial to model how the topic of the whole paper and the contexts of the citation places should affect the selection and ranking of citations. To tackle the problem, we develop a context-aware approach. The core idea is to design a novel non-parametric probabilistic model which can measure the context-based relevance between a citation context and a document. Our approach can recommend citations for a context effectively. Moreover, it can recommend a set of citations for a paper with high quality. We implement a prototype system in CiteSeerX. An extensive empirical evaluation in the CiteSeerX digital library against many baselines demonstrates the effectiveness and the scalability of our approach.

Topic detection (TD) is a fundamental research issue in the Topic Detection and Tracking (TDT) community with practical implications; TD helps analysts to separate the wheat from the chaff among the thousands of incoming news streams. In this paper, we propose a simple and effective topic detection model called the temporal Discriminative Probabilistic Model (DPM), which is shown to be theoretically equivalent to the classic vector space model with feature selection and temporally discriminative weights. We compare DPM to its various probabilistic cousins, ranging from mixture models like von-Mises Fisher (vMF) to mixed membership models like Latent Dirichlet Allocation (LDA). Benchmark results on the TDT3 data set show that sophisticated models, such as vMF and LDA, do not necessarily lead to better results; in the case of LDA, notably worst performance was obtained under variational inference, which is likely due to the significantly large number of LDA model parameters involved for document-level topic detection. On the contrary, using a relatively simple time-aware probabilistic model such as DPM suffices for both offline and online topic detection tasks, making DPM a theoretically elegant and effective model for practical topic detection.

Understanding how topics in scientific literature evolve is an interesting and important problem. Previous work simply models each paper as a bag of words and also considers the impact of authors. However, the impact of one document on another as captured by citations, one important inherent element in scientific literature, has not been considered. In this paper, we address the problem of understanding topic evolution by leveraging citations, and develop citation-aware approaches. We propose an iterative topic evolution learning framework by adapting the Latent Dirichlet Allocation model to the citation network and develop a novel inheritance topic model. We evaluate the effectiveness and efficiency of our approaches and compare with the state of the art approaches on a large collection of more than 650,000 research papers in the last 16 years and the citation network enabled by CiteSeerX. The results clearly show that citations can help to understand topic evolution better.

Web query recommendation has long been considered a key feature of search engines. Building a good Web query recommendation system, however, is very difficult due to the fundamental challenge of predicting users' search intent, especially given the limited user context information. In this paper, we propose a novel "sequential query prediction" approach that tries to grasp a user's search intent based on his/her past query sequence and its resemblance to historical query sequence models mined from massive search engine logs. Different query sequence models were examined, including the naive variable length N-gram model, variable memory Markov (VMM) model, and our proposed mixture variable memory Markov (MVMM) model. Extensive experiments were conducted to benchmark our sequence prediction algorithms against two conventional pairwise approaches on large-scale search logs extracted from a commercial search engine. Results show that the sequence-wise approaches significantly outperform the conventional pair-wise ones in terms of prediction accuracy. In particular, our MVMM approach, consistently leads the pack, making it an effective and practical approach towards Web query recommendation.

Query suggestion plays an important role in improving the usability of search engines. Although some recently proposed methods can make meaningful query suggestions by mining query patterns from search logs, none of them are context-aware - they do not take into account the immediately preceding queries as context in query suggestion. In this paper, we propose a novel context-aware query suggestion approach which is in two steps. In the offine model-learning step, to address data sparseness, queries are summarized into concepts by clustering a click-through bipartite. Then, from session data a concept sequence suffix tree is constructed as the query suggestion model. In the online query suggestion step, a user's search context is captured by mapping the query sequence submitted by the user to a sequence of concepts. By looking up the context in the concept sequence sufix tree, our approach suggests queries to the user in a context-aware manner. We test our approach on a large-scale search log of a commercial search engine containing 1:8 billion search queries, 2:6 billion clicks, and 840 million query sessions. The experimental results clearly show that our approach outperforms two baseline methods in both coverage and quality of suggestions.

Specialists who analyze online news have a hard time separating the wheat from the chaff. Moreover, automatic data-mining techniques like clustering of news streams into topical groups can fully recover the underlying true class labels of data if and only if all classes are well separated. In reality, especially for news streams, this is clearly not the case. The question to ask is thus this: if we cannot recover the full C classes by clustering, what is the largest K < C clusters we can find that best resemble the K underlying classes? Using the intuition that bursty topics are more likely to correspond to important events that are of interest to analysts, we propose several new bursty vector space models (B-VSM)for representing a news document. B-VSM takes into account the burstiness (across the full corpus and whole duration) of each constituent word in a document at the time of publication. We benchmarked our B-VSM against the classical TFIDF-VSM on the task of clustering a collection of news stream articles with known topic labels. Experimental results show that B-VSM was able to find the burstiest clusters/topics. Further, it also significantly improved the recall and precision for the top K clusters/topics.

Text representation plays a crucial role in classical text mining, where the primary focus was on static text. Nevertheless, well-studied static text representations including TFIDF are not optimized for non-stationary streams of information such as news, discussion board messages, and blogs. We therefore introduce a new temporal representation for text streams based on bursty features. Our bursty text representation differs significantly from traditional schemes in that it 1) dynamically represents documents over time, 2) amplifies a feature in proportional to its burstiness at any point in time, and 3) is topic independent. Our bursty text representation model was evaluated against a classical bag-of-words text representation on the task of clustering TDT3 topical text streams. It was shown to consistently yield more cohesive clusters in terms of cluster purity and cluster/class entropies. This new temporal bursty text representation can be extended to most text mining tasks involving a temporal dimension, such as modeling of online blog pages.

We consider the problem of analyzing word trajectories in both time and frequency domains, with the specific goal of identifying important and less-reported, periodic and aperiodic words. A set of words with identical trends can be grouped together to reconstruct an event in a completely un-supervised manner. The document frequency of each word across time is treated like a time series, where each element is the document frequency - inverse document frequency (DFIDF) score at one time point. In this paper, we 1) first applied spectral analysis to categorize features for different event characteristics: important and less-reported, periodic and aperiodic; 2) modeled aperiodic features with Gaussian density and periodic features with Gaussian mixture densities, and subsequently detected each feature's burst by the truncated Gaussian approach; 3) proposed an unsupervised greedy event detection algorithm to detect both aperiodic and periodic events. All of the above methods can be applied to time series data in general. We extensively evaluated our methods on the 1-year Reuters News Corpus [3] and showed that they were able to uncover meaningful aperiodic and periodic events.

The idea of event detection is to identify interesting patterns from a constant stream of incoming news documents. Previous research in event detection has largely focused on identifying the first event or tracking subsequent events belonging to a set of pre-assigned topics such as earthquakes, airline disasters, etc. In this paper, we propose a new problem, called Anticipatory Event Detection (AED), which aims to detect if a user-specified event has transpired. AED can be viewed as a personalized combination of event tracking and new event detection. We propose using a sentence classification approach to solve the AED problem for a restricted domain; detecting articles that describe final game results from NBA basketball news. Experimental results demonstrate the feasibility of our proposed AED solution.