arXiv subjects

Zhizhong Tan

Publications and source records attributed to Zhizhong Tan.

5 recordsLinked to original sources

arXiv2026

Credible Plan-Driven RAG Method for Multi-Hop Question Answering

Retrieval-augmented generation (RAG) has demonstrated strong performance in single-hop question answering (QA) by integrating external knowledge into large language models (LLMs). However, its effectiveness remains limited in multi-hop QA, which demands both stable reasoning and factual consistency. Existing approaches often provide partial solutions, addressing either reasoning trajectory stability or factual verification, but rarely achieving both simultaneously. To bridge this gap, we propose PAR-RAG, a three-stage Plan-then-Act-and-Review framework inspired by the PDCA cycle. PAR-RAG incorporates semantic complexity as a unifying principle through three key components: (i) complexity-aware exemplar selection guides plan generation by aligning decomposition granularity with question difficulty, thereby stabilizing reasoning trajectories; (ii) execution follows a structured retrieve-then-read process; and (iii) dual verification identifies and corrects intermediate errors while dynamically adjusting verification strength based on question complexity: emphasizing accuracy for simple queries and multi-evidence consistency for complex ones. This cognitively inspired framework integrates theoretical grounding with practical robustness. Experiments across diverse benchmarks demonstrate that PAR-RAG consistently outperforms competitive baselines, while ablation studies confirm the complementary roles of complexity-aware planning and dual verification. Collectively, these results establish PAR-RAG as a robust and generalizable framework for reliable multi-hop reasoning.

cs.CL↗

arXiv2026

HiMem: Hierarchical Long-Term Memory for LLM Long-Horizon Agents

Although long-term memory systems have made substantial progress in recent years, they still exhibit clear limitations in adaptability, scalability, and self-evolution under continuous interaction settings. Inspired by cognitive theories, we propose HiMem, a hierarchical long-term memory framework for long-horizon dialogues, designed to support memory construction, retrieval, and dynamic updating during sustained interactions. HiMem constructs cognitively consistent Episode Memory via a Topic-Aware Event--Surprise Dual-Channel Segmentation strategy, and builds Note Memory that captures stable knowledge through a multi-stage information extraction pipeline. These two memory types are semantically linked to form a hierarchical structure that bridges concrete interaction events and abstract knowledge, enabling efficient retrieval without sacrificing information fidelity. HiMem supports both hybrid and best-effort retrieval strategies to balance accuracy and efficiency, and incorporates conflict-aware Memory Reconsolidation to revise and supplement stored knowledge based on retrieval feedback. This design enables continual memory self-evolution over long-term use. Experimental results on long-horizon dialogue benchmarks demonstrate that HiMem consistently outperforms representative baselines in accuracy, consistency, and long-term reasoning, while maintaining favorable efficiency. Overall, HiMem provides a principled and scalable design paradigm for building adaptive and self-evolving LLM-based conversational agents. The code is available at https://github.com/jojopdq/HiMem.

cs.AI↗

arXiv2025

FedGRec: Dynamic Spatio-Temporal Federated Graph Learning for Secure and Efficient Cross-Border Recommendations

Due to the highly sensitive nature of certain data in cross-border sharing, collaborative cross-border recommendations and data sharing are often subject to stringent privacy protection regulations, resulting in insufficient data for model training. Consequently, achieving efficient cross-border business recommendations while ensuring privacy security poses a significant challenge. Although federated learning has demonstrated broad potential in collaborative training without exposing raw data, most existing federated learning-based GNN training methods still rely on federated averaging strategies, which perform suboptimally on highly heterogeneous graph data. To address this issue, we propose FedGRec, a privacy-preserving federated graph learning method for cross-border recommendations. FedGRec captures user preferences from distributed multi-domain data to enhance recommendation performance across all domains without privacy leakage. Specifically, FedGRec leverages collaborative signals from local subgraphs associated with users or items to enrich their representation learning. Additionally, it employs dynamic spatiotemporal modeling to integrate global and local user preferences in real time based on business recommendation states, thereby deriving the final representations of target users and candidate items. By automatically filtering relevant behaviors, FedGRec effectively mitigates noise interference from unreliable neighbors. Furthermore, through a personalized federated aggregation strategy, FedGRec adapts global preferences to heterogeneous domain data, enabling collaborative learning of user preferences across multiple domains. Extensive experiments on three datasets demonstrate that FedGRec consistently outperforms competitive single-domain and cross-domain baselines while effectively preserving data privacy in cross-border recommendations.

cs.LG↗

arXiv2025

Graph Privacy: A Heterogeneous Federated GNN for Trans-Border Financial Data Circulation

The sharing of external data has become a strong demand of financial institutions, but the privacy issue has led to the difficulty of interconnecting different platforms and the low degree of data openness. To effectively solve the privacy problem of financial data in trans-border flow and sharing, to ensure that the data is available but not visible, to realize the joint portrait of all kinds of heterogeneous data of business organizations in different industries, we propose a Heterogeneous Federated Graph Neural Network (HFGNN) approach. In this method, the distribution of heterogeneous business data of trans-border organizations is taken as subgraphs, and the sharing and circulation process among subgraphs is constructed as a statistically heterogeneous global graph through a central server. Each subgraph learns the corresponding personalized service model through local training to select and update the relevant subset of subgraphs with aggregated parameters, and effectively separates and combines topological and feature information among subgraphs. Finally, our simulation experimental results show that the proposed method has higher accuracy performance and faster convergence speed than existing methods.

cs.LG↗

arXiv2023

Futures Quantitative Investment with Heterogeneous Continual Graph Neural Network

This study aims to address the challenges of futures price prediction in high-frequency trading (HFT) by proposing a continuous learning factor predictor based on graph neural networks. The model integrates multi-factor pricing theories with real-time market dynamics, effectively bypassing the limitations of existing methods that lack financial theory guidance and ignore various trend signals and their interactions. We propose three heterogeneous tasks, including price moving average regression, price gap regression and change-point detection to trace the short-, intermediate-, and long-term trend factors present in the data. In addition, this study also considers the cross-sectional correlation characteristics of future contracts, where prices of different futures often show strong dynamic correlations. Each variable (future contract) depends not only on its historical values (temporal) but also on the observation of other variables (cross-sectional). To capture these dynamic relationships more accurately, we resort to the spatio-temporal graph neural network (STGNN) to enhance the predictive power of the model. The model employs a continuous learning strategy to simultaneously consider these tasks (factors). Additionally, due to the heterogeneity of the tasks, we propose to calculate parameter importance with mutual information between original observations and the extracted features to mitigate the catastrophic forgetting (CF) problem. Empirical tests on 49 commodity futures in China's futures market demonstrate that the proposed model outperforms other state-of-the-art models in terms of prediction accuracy. Not only does this research promote the integration of financial theory and deep learning, but it also provides a scientific basis for actual trading decisions.

cs.LG↗