Current Large Language Models excel in general reasoning yet struggle with specialized tasks requiring proprietary or domain-specific knowledge. Fine-tuning large models for every niche application is often infeasible due to black-box constraints and high computational overhead. To address this, we propose a collaborative framework that pairs a specialized weak model with a general strong model. The weak model, tailored to specific domains, produces initial drafts and background information, while the strong model leverages its advanced reasoning to refine these drafts, extending LLMs’ capabilities to critical yet specialized tasks. To optimize this collaboration, we introduce a collaborative feedback to fine-tunes the weak model, which quantifies the influence of the weak model’s contributions in the collaboration procedure and establishes preference pairs to guide preference tuning of the weak model. We validate our framework through experiments on three domains. We find that the collaboration significantly outperforms each model alone by leveraging complementary strengths. Moreover, aligning the weak model with the collaborative preference further enhances overall performance.

State-of-the-art automatic event detection struggles with interpretability and adaptability to evolving large-scale key events-unlike episodic structures, which excel in these areas.Often overlooked, episodes represent cohesive clusters of core entities performing actions at a specific time and location; a partially ordered sequence of episodes can represent a key event.This paper introduces a novel task, episode detection, which identifies episodes within a news corpus of key event articles.Detecting episodes poses unique challenges, as they lack explicit temporal or locational markers and cannot be merged using semantic similarity alone.While large language models (LLMs) can aid with these reasoning difficulties, they suffer with long contexts typical of news corpora.To address these challenges, we introduce EpiMine, an unsupervised framework that identifies a key event's candidate episodes by leveraging natural episodic partitions in articles, estimated through shifts in discriminative term combinations.These candidate episodes are more cohesive and representative of true episodes, synergizing with LLMs to better interpret and refine them into final episodes.We apply EpiMine to our three diverse, real-world event datasets annotated at the episode level, where it achieves a 59.2% average gain across all metrics compared to baselines.

Node classification is a fundamental task in graph analysis, with broad applications across various fields. Recent breakthroughs in Large Language Models (LLMs) have enabled LLM-based approaches for this task. Although many studies demonstrate the impressive performance of LLM-based methods, the lack of clear design guidelines may hinder their practical application. In this work, we aim to establish such guidelines through a fair and systematic comparison of these algorithms. As a first step, we developed LLMNodeBed, a comprehensive codebase and testbed for node classification using LLMs. It includes 10 homophilic datasets, 4 heterophilic datasets, 8 LLM-based algorithms, 8 classic baselines, and 3 learning paradigms. Subsequently, we conducted extensive experiments, training and evaluating over 2,700 models, to determine the key settings (e.g., learning paradigms and homophily) and components (e.g., model size and prompt) that affect performance. Our findings uncover 8 insights, e.g., (1) LLM-based methods can significantly outperform traditional methods in a semi-supervised setting, while the advantage is marginal in a supervised setting; (2) Graph Foundation Models can beat open-source LLMs but still fall short of strong LLMs like GPT-4o in a zero-shot setting. We hope that the release of LLMNodeBed, along with our insights, will facilitate reproducible research and inspire future studies in this field. Codes and datasets are released at https://llmnodebed.github.io/.

Large Language Models (LLMs) have revolutionized natural language processing with their remarkable capabilities in text generation and reasoning. However, these models face critical challenges when deployed in real-world applications, including hallucination generation, outdated knowledge, and limited domain expertise. Retrieval And Structuring (RAS) Augmented Generation addresses these limitations by integrating dynamic information retrieval with structured knowledge representations. This survey (1) examines retrieval mechanisms including sparse, dense, and hybrid approaches for accessing external knowledge; (2) explore text structuring techniques such as taxonomy construction, hierarchical classification, and information extraction that transform unstructured text into organized representations; and (3) investigate how these structured representations integrate with LLMs through prompt-based methods, reasoning frameworks, and knowledge embedding techniques. It also identifies technical challenges in retrieval efficiency, structure quality, and knowledge integration, while highlighting research opportunities in multimodal retrieval, cross-lingual structures, and interactive systems. This comprehensive overview provides researchers and practitioners with insights into RAS methods, applications, and future directions.

Targeted community detection aims to distinguish a particular type of community in the network. This is an important task with a lot of real-world applications, e.g., identifying fraud groups in transaction networks. Traditional community detection methods fail to capture the specific features of the targeted community and detect all types of communities indiscriminately. Semi-supervised community detection algorithms, emerged as a feasible alternative, are inherently constrained by their limited adaptability and substantial reliance on a large amount of labeled data, which demands extensive domain knowledge and manual effort.

Massive amount of unstructured text data are generated daily, ranging from news articles to scientific papers. How to mine structured knowledge from the text data remains a crucial research question. Recently, large language models (LLMs) have shed light on the text mining field with their superior text understanding and instruction-following ability. There are typically two ways of utilizing LLMs: fine-tune the LLMs with human-annotated training data, which is labor intensive and hard to scale; prompt the LLMs in a zero-shot or few-shot way, which cannot take advantage of the useful information in the massive text data. Therefore, it remains a challenge on automated mining of structured knowledge from massive text data in the era of large language models.

Understanding documents with rich layouts and multi-modal components is a long-standing and practical task. Recent Large Vision-Language Models (LVLMs) have made remarkable strides in various tasks, particularly in single-page document understanding (DU). However, their abilities on long-context DU remain an open problem. This work presents MMLONGBENCH-DOC, a long-context, multi- modal benchmark comprising 1,082 expert-annotated questions. Distinct from previous datasets, it is constructed upon 135 lengthy PDF-formatted documents with an average of 47.5 pages and 21,214 textual tokens. Towards comprehensive evaluation, answers to these questions rely on pieces of evidence from (1) different sources (text, image, chart, table, and layout structure) and (2) various locations (i.e., page number). Moreover, 33.7\% of the questions are cross-page questions requiring evidence across multiple pages. 20.6\% of the questions are designed to be unanswerable for detecting potential hallucinations. Experiments on 14 LVLMs demonstrate that long-context DU greatly challenges current models. Notably, the best-performing model, GPT-4o, achieves an F1 score of only 44.9\%, while the second-best, GPT-4V, scores 30.5\%. Furthermore, 12 LVLMs (all except GPT-4o and GPT-4V) even present worse performance than their LLM counterparts which are fed with lossy-parsed OCR documents. These results validate the necessity of future research toward more capable long-context LVLMs.

Online groups have become increasingly prevalent, providing users with space to share experiences and explore interests. Therefore, user-centric group discovery task, i.e., recommending groups to users can help both users' online experiences and platforms' long-term developments. Existing recommender methods can not deal with this task as modeling user-group participation into a bipartite graph overlooks their item-side interests. Although there exist a few works attempting to address this task, they still fall short in fully preserving the social context and ensuring effective interest representation learning.

Temporal Graph Networks (TGNs) have shown remarkable performance in learning representation for continuous-time dynamic graphs. However, real-world dynamic graphs typically contain diverse and intricate noise. Noise can significantly degrade the quality of representation generation, impeding the effectiveness of TGNs in downstream tasks. Though structure learning is widely applied to mitigate noise in static graphs, its adaptation to dynamic graph settings poses two significant challenges. i) Noise dynamics. Existing structure learning methods are ill-equipped to address the temporal aspect of noise, hampering their effectiveness in such dynamic and ever-changing noise patterns. ii) More severe noise. Noise may be introduced along with multiple interactions between two nodes, leading to the re-pollution of these nodes and consequently causing more severe noise compared to static graphs.

Large language models with instructionfollowing capabilities open the door to a wider group of users.However, when it comes to information extraction -a classic task in natural language processing -most task-specific systems cannot align well with long-tail ad hoc extraction use cases for non-expert users.To address this, we propose a novel paradigm, termed On-Demand Information Extraction, to fulfill the personalized demands of real-world users.Our task aims to follow the instructions to extract the desired content from the associated text and present it in a structured tabular format.The table headers can either be userspecified or inferred contextually by the model.To facilitate research in this emerging area, we present a benchmark named INSTRUCTIE, inclusive of both automatically generated training data, as well as the human-annotated test set.Building on INSTRUCTIE, we further develop an On-Demand Information Extractor, ODIE.Comprehensive evaluations on our benchmark reveal that ODIE substantially outperforms the existing open-source models of similar size.Our code and dataset are released on https://github.com/yzjiao/On-Demand-IE.

Siru Ouyang, Shuohang Wang, Yang Liu, Ming Zhong, Yizhu Jiao, Dan Iter, Reid Pryzant, Chenguang Zhu, Heng Ji, Jiawei Han. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.

Massive and fast-evolving news articles keep emerging on the web. To effectively summarize and provide concise insights into real-world events, we propose a new event knowledge extraction task Event Chain Mining in this paper. Given multiple documents about a super event, it aims to mine a series of salient events in temporal order. For example, the event chain of super event Mexico Earthquake in 2017 is {earthquake hit Mexico, destroy houses, kill people, block roads}. This task can help readers capture the gist of texts quickly, thereby improving reading efficiency and deepening text comprehension. To address this task, we regard an event as a cluster of different mentions of similar meanings. In this way, we can identify the different expressions of events, enrich their semantic knowledge and replenish relation information among them. Taking events as the basic unit, we present a novel unsupervised framework, EMiner. Specifically, we extract event mentions from texts and merge them with similar meanings into a cluster as a single event. By jointly incorporating both content and commonsense, essential events are then selected and arranged chronologically to form an event chain. Meanwhile, we annotate a multi-document benchmark to build a comprehensive testbed for the proposed task. Extensive experiments are conducted to verify the effectiveness of EMiner in terms of both automatic and human evaluations.

Since group activities have become very common in daily life, there is an urgent demand for generating recommendations for a group of users, referred to as group recommendation task. Existing group recommendation methods usually infer groups’ preferences via aggregating diverse members’ interests. Actually, groups’ ultimate choice involves compromises between members, and finally, an agreement can be reached. However, existing individual information aggregation lacks a holistic group-level consideration, failing to capture the consensus information. Besides, their specific aggregation strategies either suffer from high computational costs or become too coarse-grained to make precise predictions.

Multi-dimensional evaluation is the dominant paradigm for human evaluation in Natural Language Generation (NLG), i.e., evaluating the generated text from multiple explainable dimensions, such as coherence and fluency.However, automatic evaluation in NLG is still dominated by similarity-based metrics, and we lack a reliable framework for a more comprehensive evaluation of advanced models.In this paper, we propose a unified multi-dimensional evaluator UNIEVAL for NLG.We re-frame NLG evaluation as a Boolean Question Answering (QA) task, and by guiding the model with different questions, we can use one evaluator to evaluate from multiple dimensions.Furthermore, thanks to the unified Boolean QA format, we are able to introduce an intermediate learning phase that enables UNIEVAL to incorporate external knowledge from multiple related tasks and gain further improvement.Experiments on three typical NLG tasks show that UNIEVAL correlates substantially better with human judgments than existing metrics.Specifically, compared to the top-performing unified evaluators, UNIEVAL achieves a 23% higher correlation on text summarization, and over 43% on dialogue response generation.Also, UNIEVAL demonstrates a strong zero-shot learning ability for unseen evaluation dimensions and tasks.Source code, data and all pre-trained evaluators are available on our GitHub repository 1 . Generated Summary:Harry Kane is nominated for both the PFA player and young player of the season.The Spurs striker has been released from the awards ceremony on Sunday.The Tottenham striker features in a new animation.Reference Summary: Harry Kane has been in superb form for Tottenham this season.The 21-year-old has scored 30 goals in all competitions for Spurs.Kane also made his England debut and scored within two minutes.Document: Harry Kane's celebrations this season have always shown him to be an animated young man . . .Similarity-based Evaluators ROUGE-1: 0.44 ROUGE-2: 0.25 ROUGE-L: 0.42 BERTScore: 0.24 Single-dimensional Evaluators (predicted by two different evaluators (Deng et al., 2021)) Consistency: 0.87 Relevance: 0.74 Unified Evaluator (predicted by BARTScore, and the scoring range is negative infinity to 0) Precision: -5.45 Recall: -4.93 F1: -5.19

Community detection refers to the task of discovering closely related subgraphs to understand the networks. However, traditional community detection algorithms fail to pinpoint a particular kind of community. This limits its applicability in real-world networks, e.g., distinguishing fraud groups from normal ones in transaction networks. Recently, semi-supervised community detection emerges as a solution. It aims to seek other similar communities in the network with few labeled communities as training data. Existing works can be regarded as seed-based: locate seed nodes and then develop communities around seeds. However, these methods are quite sensitive to the quality of selected seeds since communities generated around a mis-detected seed may be irrelevant. Besides, they have individual issues, e.g., inflexibility and high computational overhead. To address these issues, we propose CLARE, which consists of two key components, Community Locator and Community Rewriter. Our idea is that we can locate potential communities and then refine them. Therefore, the community locator is proposed for quickly locating potential communities by seeking subgraphs that are similar to training ones in the network. To further adjust these located communities, we devise the community rewriter. Enhanced by deep reinforcement learning, it suggests intelligent decisions, such as adding or dropping nodes, to refine community structures flexibly. Extensive experiments verify both the effectiveness and efficiency of our work compared with prior state-of-the-art approaches on multiple real-world datasets.

Click-through rate prediction is a critical task in online advertising. Currently, many existing methods attempt to extract user potential interests from historical click behavior sequences. However, it is difficult to handle sparse user behaviors or broaden interest exploration. Recently, some researchers incorporate the item-item co-occurrence graph as an auxiliary. Due to the elusiveness of user interests, those works still fail to determine the real motivation of user click behaviors. Besides, those works are more biased towards popular or similar commodities. They lack an effective mechanism to break the diversity restrictions. In this paper, we point out two special properties of triangles in the item-item graphs for recommendation systems: Intra-triangle homophily and Inter-triangle heterophiy. Based on this, we propose a novel and effective framework named Triangle Graph Interest Network (TGIN). For each clicked item in user behavior sequences, we introduce the triangles in its neighborhood of the item-item graphs as a supplement. TGIN regards these triangles as the basic units of user interests, which provide the clues to capture the real motivation for a user clicking an item. We characterize every click behavior by aggregating the information of several interest units to alleviate the elusive motivation problem. The attention mechanism determines users' preference for different interest units. By selecting diverse and relative triangles, \short brings in novel and serendipitous items to expand exploration opportunities of user interests. Then, we aggregate the multi-level interests of historical behavior sequences to improve CTR prediction. Extensive experiments on both of public and industrial datasets clearly verify the effectiveness of our framework.

Graph Neural Networks(GNNs), like GCN and GAT, have achieved great success in a number of supervised or semi-supervised tasks including node classification and link prediction. These existing graph neural networks can effectively encode neighborhood information of graph nodes through their message aggregating mechanisms. However, there are some unsupervised and structure-related tasks like community detection, which is a fundamental problem in network analysis that finds densely-connected groups of nodes and separates them from others in graphs. It is still difficult for these general-purposed GNNs to learn the needed structural information in these particular problems. To overcome the shortcomings of general-purposed graph representation learning methods, we propose the Community Deep Graph Infomax (CommDGI), a graph neural network designed to handle community detection problems. Inspired by the success of deep graph infomax in self-supervised graph learning, we design a novel mutual information mechanism to capture neighborhood as well as community information in graphs. A trainable clustering layer is employed to learn the community partition in an end-to-end manner. Disentangled representation learning is applied in our graph neural network so that the model can improve interpretability and generalization. Throughout the whole learning process, joint optimization is applied to learn the community-related node representations. The experimental results show that our algorithm outperforms state-of-the-art community detection methods.

Graph representation learning has attracted lots of attention recently. Existing graph neural networks fed with the complete graph data are not scalable due to limited computation and memory costs. Thus, it remains a great challenge to capture rich information in large-scale graph data. Besides, these methods mainly focus on supervised learning and highly depend on node label information, which is expensive to obtain in the real world. As to unsupervised network embedding approaches, they overemphasize node proximity instead, whose learned representations can hardly be used in downstream application tasks directly. In recent years, emerging self-supervised learning provides a potential solution to address the aforementioned problems. However, existing self-supervised works also operate on the complete graph data and are biased to fit either global or very local (1-hop neighborhood) graph structures in defining the mutual information based loss terms. In this paper, a novel self-supervised representation learning method via Sub-graph Contrast, namely Subg-Con, is proposed by utilizing the strong correlation between central nodes and their sampled subgraphs to capture regional structure information. Instead of learning on the complete input graph data, with a novel data augmentation strategy, Subg-Con learns node representations through a contrastive loss defined based on subgraphs sampled from the original graph instead. Compared with existing graph representation learning approaches, Subg-Con has prominent performance advantages in weaker supervision requirements, model learning scalability, and parallelization. Extensive experiments verify both the effectiveness and the efficiency of our work compared with both classic and state-of-the-art graph representation learning approaches on multiple realworld large-scale benchmark datasets from different domains.

To enjoy more social network services, users nowadays are usually involved in multiple online sites at the same time. Aligned social networks provide more information to alleviate the problem of data insufficiency. In this paper, we target on the collective link prediction problem and aim to predict both the intra-network social links as well as the inter-network anchor links across multiple aligned social networks. It is not an easy task, and the major challenges involve the network characteristic difference problem and different directivity properties of the social and anchor links to be predicted. To address the problem, we propose an application oriented network embedding framework, Hierarchical Graph Attention based Network Embedding (HGANE), for collective link prediction over directed aligned networks. Very different from the conventional general network embedding models, HGANE effectively incorporates the collective link prediction task objectives into consideration. It learns the representations of nodes by aggregating information from both the intra-network neighbors (connected by social links) and inter-network partners (connected by anchor links). What's more, we introduce a hierarchical graph attention mechanism for the intra-network neighbors and inter-network partners respectively, which resolves the network characteristic differences and the link directivity challenges effectively. Extensive experiments have been conducted on real-world aligned networks datasets to demonstrate that our model outperformed the state-of-the-art baseline methods in addressing the collective link prediction problem by a large margin.