Information Processing & Management最新文献

Empathetic response generation is a crucial task in natural language processing, enabling emotionally resonant machine–human interactions. In this paper, we introduce the InfRa (Integrating Discourse Features and Response Assessment) model to address limitations in traditional methods for this task, such as the lack of deep dialogue comprehension and response control. InfRa integrates discourse features to augment structural dialogue understanding, with a novel edge pruning and mutual information learning module to further refine the representation. The model also employs a response evaluation module for dynamic optimization, ensuring emotional and semantic consistency between the generated response and its context. Our experiments demonstrate that InfRa outperforms existing baselines, reducing the Perplexity (PPL) score by approximately 9 points and excelling in all three fine-grained aspects of human evaluation. This research not only advances the development of empathetic chatbots but also provides valuable insights for broader text generation tasks.

Common law courts need to refer to similar precedents’ judgments to inform their current decisions. Generating high-quality summaries of court judgment documents can facilitate legal practitioners to efficiently review previous cases and assist the general public in accessing how the courts operate and how the law is applied. Previous court judgment summarization research focuses on civil law or a particular jurisdiction’s judgments. However, judges can refer to the judgments from all common law jurisdictions. Current summarization datasets are insufficient to satisfy the demands of summarizing precedents across multiple jurisdictions, especially when labeled data are scarce for many jurisdictions. To address the lack of datasets, we present CLSum, the first dataset for summarizing multi-jurisdictional common law court judgment documents. Besides, this is the first court judgment summarization work adopting large language models (LLMs) in data augmentation, summary generation, and evaluation. Specifically, we design an LLM-based data augmentation method incorporating legal knowledge. We also propose a legal knowledge enhanced evaluation metric based on LLM to assess the quality of generated judgment summaries. Our experimental results verify that the LLM-based summarization methods can perform well in the few-shot and zero-shot settings. Our LLM-based data augmentation method can mitigate the impact of low data resources. Furthermore, we carry out comprehensive comparative experiments to find essential model components and settings that are capable of enhancing summarization performance.

As one of the typical formats of prevalent user-generated content in social media platforms, micro-videos inherently incorporate multimodal characteristics associated with a group of label concepts. However, existing methods generally explore the consensus features aggregated from all modalities to train a final multi-label predictor, while overlooking fine-grained semantic dependencies between modality and label domains. To address this problem, we present a novel multimodal deep hierarchical semantic-aligned matrix factorization (DHSAMF) method, which is devoted to bridging the dual-domain semantic discrepancies and the inter-modal heterogeneity gap for solving the multi-label classification task of micro-videos. Specifically, we utilize deep matrix factorization to individually explore the hierarchical modality-specific representations. A series of semantic embeddings is introduced to facilitate latent semantic interactions between modality-specific representations and label features in a layerwise manner. To further improve the representation ability of each modality, we leverage underlying correlation structures among instances to adequately mine intra-modal complementary attributes, and maximize the inter-modal alignment by aggregating consensus attributes in an optimal permutation. The experimental results conducted on the MTSVRC and VidOR datasets have demonstrated that our DHSAMF outperforms other state-of-the-art methods by nearly 3% and 4% improvements in terms of the AP metric.

Users rely on exploratory search to find useful and serendipitous information in online knowledge communities. Although there are multiple types of exploratory tasks, we know little about the differences in search behaviors for distinct exploratory tasks. Consequently, communities cannot provide adaptive support for users performing distinct exploratory tasks. Against this backdrop, a lab experiment was conducted to reveal the behavioral differences among different exploratory tasks through querying, clicking, scrolling and eye-tracking data. By operationalizing search motivation and cognitive complexity, exploratory tasks were categorized into four types: borderline learning, core learning, borderline investigation, and core investigation. 37 participants with good search ability completed the experiment, and the final dataset contains 124 observations from 31 participants. ANOVA tests showed that users performing investigation tasks generated longer queries, more satisfied clicks, less scrolling, more fixations within result areas, more interactions with social tags, and more frequent browsing of reviews than users performing learning tasks. Compared to core tasks, users had more queries when performing borderline tasks. Moreover, machine learning was conducted to validate whether different exploratory tasks can be distinguished through these behaviors. Gradient Boosting Machine allowed the correct classification of four exploratory tasks with 84.75 % accuracy. The three most important indicators were UniQueryNum, MaxScrollDepth, and TagClickNum. By revealing differences in user behaviors for different exploratory tasks, this study advances the understanding of exploratory search behavior in knowledge communities at a finer granularity, and helps develop adaptive communities that support distinct exploratory tasks.

There is an important ongoing effort aimed to tackle misinformation and to perform reliable fact-checking by employing human assessors at scale, with a crowdsourcing-based approach. Previous studies on the feasibility of employing crowdsourcing for the task of misinformation detection have provided inconsistent results: some of them seem to confirm the effectiveness of crowdsourcing for assessing the truthfulness of statements and claims, whereas others fail to reach an effectiveness level higher than automatic machine learning approaches, which are still unsatisfactory. In this paper, we aim at addressing such inconsistency and understand if truthfulness assessment can indeed be crowdsourced effectively. To do so, we build on top of previous studies; we select some of those reporting low effectiveness levels, we highlight their potential limitations, and we then reproduce their work attempting to improve their setup to address those limitations. We employ various approaches, data quality levels, and agreement measures to assess the reliability of crowd workers when assessing the truthfulness of (mis)information. Furthermore, we explore different worker features and compare the results obtained with different crowds. According to our findings, crowdsourcing can be used as an effective methodology to tackle misinformation at scale. When compared to previous studies, our results indicate that a significantly higher agreement between crowd workers and experts can be obtained by using a different, higher-quality, crowdsourcing platform and by improving the design of the crowdsourcing task. Also, we find differences concerning task and worker features and how workers provide truthfulness assessments.

Static index pruning aims to remove redundant parts of an index to reduce the file size and query processing time. In this paper, we focus on the impact of index pruning on the topical diversity of query results obtained over these pruned indexes, due to the emergence of diversity as an important metric of quality in modern search systems. We hypothesize that typical index pruning strategies are likely to harm result diversity, as the latter dimension has been vastly overlooked while designing and evaluating such methods. As a remedy, we introduce three novel diversity-aware pruning strategies aimed at maintaining the diversity effectiveness of query results. In addition to other widely used features, our strategies exploit document clustering methods and word-embeddings to assess the possible impact of index elements on the topical diversity, and to guide the pruning process accordingly. Our thorough experimental evaluations verify that typical index pruning strategies lead to a substantial decline (i.e., up to 50% for some metrics) in the diversity of the results obtained over the pruned indexes. Our diversity-aware approaches remedy such losses to a great extent, and yield more diverse query results, for which scores of the various diversity metrics are closer to those obtained over the full index. Specifically, our best-performing strategy provides gains in result diversity reaching up to 2.9%, 3.0%, 7.5%, and 3.9% wrt. the strongest baseline, in terms of the ERR-IA, $\alpha$-nDCG, P-IA, and ST-Recall metrics (at the cut-off value of 20), respectively. The proposed strategies also yield better scores in terms of an entropy-based fairness metric, confirming the correlation between topical diversity and fairness in this setup.

The recent advanced Contrastive Neural Topic Model (CNTM) was proposed to tackle topic collapse through document-level contrastive learning. However, limited by its usage of the Logistic-Normal prior in topic space and document level contrastive learning, it is less capable of disentangling semantically similar topics. To address the limitation, we propose a novel Dual Contrastive Topic Model (DCTM) that utilizes the Dirichlet prior to capture interpretable patterns. Besides, it incorporates dual (document-level and topic-level) contrastive learning on the topic distribution matrix which helps generate discriminative topic representations and mine identifiable topics. Our proposed DCTM outperforms the state-of-the-art neural topic models in terms of topic coherence and diversity, which is verified by extensive experimentation on three publicly available text corpora. In detail, the proposed DCTM surpasses baselines on almost all the used topic coherence metrics ($C_P$, $C_A$, NPMI for 20Newsgroups, $C_P$, $C_A$, NPMI and UCI for Grolier and DBPedia), and it also obtains higher topic diversity with 1 datasets respectively. Moreover, when performing text clustering, DCTM also achieves significant improvements, with observed increases of more than 1% (20Newsgroups) and 6% (DBPedia) in accuracy.

Multi-behavior recommendations effectively integrate various types of behaviors and have been proven to enhance recommendation performance. However, existing researches primarily focus on distinguishing between various behaviors, neglecting the exploration of common representations within each behavior that might reflect individual preferences from different perspectives. Meanwhile, interactions within each behavior remain sparse; how to learn effective information from limited data poses a significant challenge. In this study, we propose a simplices-based higher-order enhancement graph neural network for multi-behavior recommendations, HEM-GNN. Specifically, we adopt a supervised method to distinguish the importance of different behaviors and perform inter-behavior representation learning. Meanwhile, for each behavior, we define implicit relationships to mitigate data sparsity, and then aggregate information from nodes within simplices to extract their higher-order commonalities. Finally, HEM-GNN leverages these representations to make recommendations. Through experiments on three public datasets (Taobao, Beibei, and IJCAI), HEM-GNN demonstrates better performance compared to 10 baseline algorithms. It outperforms state-of-the-art models by margins ranging from 8.99% to 10.58% in HR@$K$ and 8.18% to 9.69% in NDCG@$K$, highlighting the significance of higher-order features in multi-behavior recommendations. The model and datasets are released at: https://github.com/SamuelZack/MultiRec.

In the rapidly evolving landscape of innovation, the early identification of emerging topics is crucial across diverse research domains. This study views weak signals as the preliminary stage of emerging topics and constructs an innovative weak signal triple-dimensional analytical framework to discern nascent emerging topics. The framework uses triads to represent signals by constructing keyword citation networks and establish a collection of novel signals through network topology analysis. Weak signals are subsequently identified by examining the visibility, diffusion and social influence of signals with time-weighted attributes. An altmetrics indicator is employed to formally measure the social influence of weak signals from the perspective of public perception. We apply the proposed framework to the field of gene editing, and the outcomes of literature analysis and dynamic validation substantiate the efficacy of our approach. Compared to related methods, our framework demonstrates a more nuanced ability to distinguish between various signals, identifying more weak signals and research topics with increased potential for social impact. This research provides valuable insights for strategic decision-making, innovation management, and future foresight.

The increase of image information brings the need for secure storage and management, and people are used to uploading images to cloud servers for storage, but the issue of privacy management and protection has become a great challenge because images may contain some sensitive information. To solve this problem, this paper proposes a novel secure and efficient fine-grained TPE scheme (FG-TPE), specifically, the image pixels are firstly divided into blocks, and multiple rounds of neighboring pixel substitution and permutation fine-grained encryption operations are performed in each block to achieve obfuscated protection of sensitive feature information of the image. Then, the state transfer process of image pixel encryption is reduction to the adversarial detection in a stochastic environment, and the optimal encryption rounds bounds are found by Kalman filtering method. Finally, experiments conducted on two face datasets show that, in qualitative and quantitative comparisons, the average encryption time is decreased remarkably, improved encryption efficiency, and the ciphertext expansion rate is reduced by 19.6% on average, possessing a better image spatiality when compared to the state-of-the-art approaches. Excellent resistance to AI restoration performance has been achieved with only 16 × 16 divided block encryption, and face detection recognition has been fully defended against 32 × 32 divided block encryption, achieving a balance between privacy security and usability management of image information.