As the cornerstone of telemedicine, Electronic Health Records (EHRs) not only reduce clinical costs but also enable precision diagnostics. As medical institutions increasingly outsource EHRs to Cloud Service Providers (CSPs), dual challenges have emerged as critical issues: preserving patient privacy and enhancing the search experience for medical personnel. While multi-keyword searchable encryption has gained significant attention in the medical domain as a potential solution, existing schemes have significant limitations in both practicality and security. First, the growing number of medical institutions complicates the management of key and privileges. Second, the impoverished search semantics in existing query mechanisms severely degrades the clinical user experience, creating unacceptable operational bottlenecks in medical practice. Furthermore, excessive reliance on CSPs leads to ignoring situations where the returned results are incorrect, impacting the availability and security of the telemedicine system. To address these limitations, we propose a Verifiable Multi-Semantic Keyword Search scheme (VMSKS) for EHRs in cloud-based telemedicine. To resolve the security requirements arising from the increasing medical institutions, we innovatively design a more efficient dual Securek-Nearest Neighbor technique (SKNN) for key distribution. Meanwhile, fine-grained access control is implemented using access policy trees, ensuring the controllability of data access. This approach safeguards the privacy of EHRs. To support flexible EHR search for medical personnel, the prime Hadamard product encoding technique is exploited to provide queries that support multiple search semantics simultaneously. Given the potential unreliability of CSPs, VMSKS introduces a novel verification mechanism by constructing verification proofs during encryption, ensuring the authenticity and integrity of returned results. Theoretical analysis and experimental evaluation demonstrate the security and efficiency of VMSKS, respectively.

Guizhen Chen, Weiwen Xu, Hao Zhang, Hou Pong Chan, Chaoqun Liu, Lidong Bing, Deli Zhao, Anh Tuan Luu, Yu Rong. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Computational argumentation has become an essential tool in various domains, including law, public policy, and artificial intelligence.It is an emerging research field in natural language processing that attracts increasing attention.Research on computational argumentation mainly involves two types of tasks: argument mining and argument generation.As large language models (LLMs) have demonstrated impressive capabilities in understanding context and generating natural language, it is worthwhile to evaluate the performance of LLMs on diverse computational argumentation tasks.This work aims to embark on an assessment of LLMs, such as ChatGPT, Flan models, and LLaMA2 models, in both zero-shot and few-shot settings.We organize existing tasks into six main categories and standardize the format of fourteen openly available datasets.In addition, we present a new benchmark dataset on counter speech generation that aims to holistically evaluate the end-to-end performance of LLMs on argument mining and argument generation.Extensive experiments show that LLMs exhibit commendable performance across most of the datasets, demonstrating their capabilities in the field of argumentation.Our analysis offers valuable suggestions for evaluating computational argumentation and its integration with LLMs in future research endeavors.1

Large language models (LLMs) have demonstrated impressive capabilities across diverse languages. This study explores how LLMs handle multilingualism. Based on observed language ratio shifts among layers and the relationships between network structures and certain capabilities, we hypothesize the LLM's multilingual workflow ($\texttt{MWork}$): LLMs initially understand the query, converting multilingual inputs into English for task-solving. In the intermediate layers, they employ English for thinking and incorporate multilingual knowledge with self-attention and feed-forward structures, respectively. In the final layers, LLMs generate responses aligned with the original language of the query. To verify $\texttt{MWork}$, we introduce Parallel Language-specific Neuron Detection ($\texttt{PLND}$) to identify activated neurons for inputs in different languages without any labeled data. Using $\texttt{PLND}$, we validate $\texttt{MWork}$ through extensive experiments involving the deactivation of language-specific neurons across various layers and structures. Moreover, $\texttt{MWork}$ allows fine-tuning of language-specific neurons with a small dataset, enhancing multilingual abilities in a specific language without compromising others. This approach results in an average improvement of $3.6\%$ for high-resource languages and $2.3\%$ for low-resource languages across all tasks with just $400$ documents.