Anonymizing text that contains sensitive information is crucial for a wide range of applications. Existing techniques face the emerging challenges of the re-identification ability of large language models (LLMs), which have shown advanced capability in memorizing detailed information and reasoning over dispersed pieces of patterns to draw conclusions. When defending against LLM-based re-identification, anonymization could jeopardize the utility of the resulting anonymized data in downstream tasks. In general, the interaction between anonymization and data utility requires a deeper understanding within the context of LLMs. In this paper, we propose a framework composed of three key LLM-based components: \textit{a privacy evaluator}, \textit{a utility evaluator} and \textit{an optimization component}, which work collaboratively to perform anonymization. Extensive experiments demonstrate that the proposed model outperforms existing baselines, showing robustness in reducing the risk of re-identification while preserving greater data utility in downstream tasks. We provide detailed studies on these core modules. To consider large-scale and real-time applications, we investigate the distillation of the anonymization capabilities into lightweight models. All of our code and datasets will be made publicly available at \texttt{[Github URL]}.

KG-to-text generation models aim to generate a natural language text with one or more sentences from input knowledge graphs (KGs). Recent models use either entity-level or word-level graphs as input data to generate texts. The Entity-level graph provides the relations between entities but neglects the relations between words within a single entity and different entities. In contrast, a word-level graph can better consider the rich semantic relations among words but can fail to capture the relations between entities. This paper proposes a novel graph-to-sequence model that encodes both entity-level and word-level KGs to enhance node representations and investigates different decoder architectures to integrate contextualized representations of words and entities effectively. Automatic and human evaluations demonstrate that our method outperforms previous single-granularity methods on two popular graph-to-text benchmarks.

Keyphrases can concisely describe the high-level topics discussed in a document that usually possesses hierarchical topic structures. Thus, it is crucial to understand the hierarchical topic structures and employ it to guide the keyphrase identification. However, integrating the hierarchical topic information into a deep keyphrase generation model is unexplored. In this paper, we focus on how to effectively exploit the hierarchical topic to improve the keyphrase generation performance (HTKG). Specifically, we propose a novel hierarchical topic-guided variational neural sequence generation method for keyphrase generation, which consists of two major modules: a neural hierarchical topic model that learns the latent topic tree across the whole corpus of documents, and a variational neural keyphrase generation model to generate keyphrases under hierarchical topic guidance. Finally, these two modules are jointly trained to help them learn complementary information from each other. To the best of our knowledge, this is the first attempt to leverage the neural hierarchical topic to guide keyphrase generation. The experimental results demonstrate that our method significantly outperforms the existing state-of-the-art methods across five benchmark datasets.