We develop a novel approach to tackle the common but challenging problem of conformal inference for missing data in machine learning, focusing on Missing at Random (MAR) data. We propose a new procedure Conformal prediction for Missing data under Multiple Robust Learning (CM–MRL) that combines split conformal calibration with a multiple robust empirical-likelihood (EL) reweighting scheme. The method proceeds via a double calibration by reweighting the complete-case scores by EL so that their distribution matches the full calibration distribution implied by MAR, even when some working models are misspecified. We demonstrate the asymptotic behavior of our estimators through empirical process theory and provide reliable coverage for our prediction intervals, both marginally and conditionally and we further show an interval-length dominance result. We show the effectiveness of the proposed method by several numerical experiments in the presence of missing data.
Technological collaboration has become an important means of sharing technological risks and advancing key technological breakthroughs. However, existing studies often overlook patent value heterogeneity in technology representation and insufficiently integrate multiple technological drivers of collaboration, while also lacking effective modeling of their dynamic evolution. This study proposes a Dynamic Multi-View Graph Neural Network model based on Semantic and Quality fusion (SQ-DMVGNN). A semantic-quality fusion module is designed to project applicants’ textual semantic features and multidimensional quality attributes into a unified high-dimensional vector space, thereby capturing both technological semantics and patent quality. Then, the proposed framework constructs a dynamic multi-view graph structures to systematically decouple and integrate technological similarity, complementarity, and strategic trend, while modeling their evolution across time slices. Through the proposed framework, SQ-DMVGNN more accurately captures the multidimensional technological relationships underlying collaborations driven by high-quality technological resources, thereby improving both the performance and interpretability of collaboration prediction in dynamic environments. This study selects the representative semiconductor industry as empirical data, obtaining a total of 668,790 non-duplicated invention patents. Experimental results demonstrate that integrating patent quality into technological portfolios enables a more precise differentiation of technological heterogeneity among organizations. Compared with baseline methods grounded in different principles, the proposed model generally achieves near-optimal performance across major evaluation metrics, particularly in AUC, AP, MRR, and Hits@K. Ablation studies not only validate the effectiveness of each auxiliary view, temporal module, and patent quality, but also investigate model substitutions. Moreover, our work conducts parameter sensitivity analyses across different modules, verifying the stability of technological knowledge extraction and collaboration prediction.
The quality of dataset annotations is crucial for Named Entity Recognition (NER), but increasing task complexity exacerbates label noise. While previous studies have made progress in addressing isolated aspects of noisy-label problems, their narrow focus limits real-world applicability. We propose a multi-label dynamically robust loss function that adapts to different noise levels manifesting during disparate training epochs through our dynamically robust manner and dynamically adapting scheme, thereby optimizing the suppression of potential noise from various noise types. We conduct extensive experiments on 10 benchmark datasets of two NER paradigms: 6 standard datasets for flat and nested NER, and 4 real-world single-label noisy datasets from NoiseBench. On standard datasets, our method achieves results that are superior or comparable to existing state-of-the-art baselines, with an average F1-score improvement of 0.2%. On these noisy datasets, we observe more significant gains: an average F1-score improvement of 0.62%, with the maximum improvement reaching 1.32%. Additionally, we evaluate our method's multi-label architectural robustness under synthetic noise on ACE05, establishing its effectiveness in such settings, while robustness under real-world multi-label noise remains to be assessed when corresponding benchmarks become available. Finally, we provide a systematic analysis of the difference between real-world and synthetic noisy labels, revealing that the current noisy-label learning in NER is facing challenges brought about by the lack of high quality noisy datasets and complex task paradigms1.
While destinations increasingly leverage cultural capital for competitiveness, research predominantly focuses on top-down promotions of monument-centered heritage, neglecting the distinct social logic of gastronomic culture. Grounded in cultural capital theory, this study aims to examine how Hong Kong's tea restaurants food culture translates into residents' recommendation intentions toward culturally diverse individuals. It untangles the core inquiries: how everyday food culture fosters attachment; how this attachment subsequently shapes cultural identity and confidence; and how nostalgia and knowledge moderate these psychological pathways. Additionally, it identifies which specific tea restaurant's food culture dimensions serve as the primary drivers. Utilizing an empirical sample of 615 Hong Kong residents, this research adopts a PLS–SEM–ANN approach. The PLS–SEM analysis confirms that cultural identity and confidence act as pivotal mediators, while the complementary ANN analysis reveals that gastronomic identity, exceeding mere sensory flavour, is the predominant driving dimension. Empirically, nostalgia amplifies the culture–attachment link, and knowledge strengthens the attachment–confidence relationship. These findings advocate for a paradigm shift in destination marketing: beyond institutional campaigns, mobilizing residents as active cultural transmitters is crucial for gastronomic tourism, establishing a scalable mechanism for everyday cultural capital transmission.
Accelerating the energy transition is pivotal to achieving sustainable economic development, yet localized environmental policies can profoundly reshape the spatial allocation of corporate capital. The impact of place-based energy transition policies on firms' cross-regional investment remains largely unexplored. Existing studies primarily examine the environmental and economic consequences of such policies, while paying limited attention to their influence on firms' geographic expansion strategies. Using a sample of Chinese A-share listed firms from 2006 to 2022, we examine whether and how the New Energy Demonstration City (NEDC) pilot policy affects corporate cross-regional investment (CORI). We find that the NEDC policy significantly increases firms' CORI, indicating that firms located in pilot cities expand more actively beyond their home provinces. This finding is consistent with the resource-based view. We further find that this effect is stronger for non-state-owned firms, non-heavy-polluting firms, and firms audited by the Big Four, and remains robust to a series of robustness and endogeneity tests. Mechanism analyses show that the NEDC policy promotes CORI by reducing information asymmetry and lowering corporate risk-taking. These findings suggest that place-based energy transition policies reshape firms’ geographic allocation of resources and provide new evidence on how policy-induced environmental change affects corporate investment decisions.
