Continual Semantic Parsing (CSP) aims to train parsers to convert natural language questions into SQL across tasks with limited annotated examples, adapting to dynamically updated databases in real-world scenarios. Previous studies mitigate this challenge by replaying historical data or employing parameter-efficient tuning (PET), but they often violate data privacy or rely on ideal continual learning settings. To address these issues, we propose a new Large Language Model (LLM)-Enhanced Continuous Semantic Parsing method, named LECSP, which alleviates forgetting while encouraging generalization, without requiring real data replay or ideal settings. Specifically, it first analyzes the commonalities and differences between tasks from the SQL syntax perspective to guide LLMs in reconstructing key memories and improving memory accuracy through calibration. Then, it uses a task-aware dual-teacher distillation framework to promote the accumulation and transfer of knowledge during sequential training. Experimental results on two CSP benchmarks show that our method significantly outperforms existing methods, even those utilizing data replay or ideal settings. Additionally, we achieve generalization performance beyond upper limits, better adapting to unseen tasks.

Continual Semantic Parsing (CSP) enables parsers to generate SQL from natural language questions in task streams, using minimal annotated data to handle dynamically evolving databases in real-world scenarios. Previous works often rely on replaying historical data, which poses privacy concerns. Recently, replay-free continual learning methods based on Parameter-Efficient Tuning (PET) have gained widespread attention. However, they often rely on ideal settings and initial task data, sacrificing the model’s generalization ability, which limits their applicability in real-world scenarios. To address this, we propose a novel Adaptive PET eXpert meta-learning (APEX) approach for CSP. First, SQL syntax guides the LLM to assist experts in adaptively warming up, ensuring better model initialization. Then, a dynamically expanding expert pool stores knowledge and explores the relationship between experts and instances. Finally, a selection/fusion inference strategy based on sample historical visibility promotes expert collaboration. Experiments on two CSP benchmarks show that our method achieves superior performance without data replay or ideal settings, effectively handling cold start scenarios and generalizing to unseen tasks, even surpassing performance upper bounds.

Quantifying the complexity of quantum states is a longstanding key problem in various subfields of science, ranging from quantum computing to the black-hole theory. The lower bound on quantum pure state complexity has been shown to grow linearly with system size [J. Haferkamp et al., 2022, *Nat. Phys.*]. However, extending this result to noisy circuit environments, which better reflect real quantum devices, remains an open challenge. In this paper, we explore the complexity of weakly noisy quantum states via the quantum learning method. We present an efficient learning algorithm, that leverages the classical shadow representation of target quantum states, to predict the circuit complexity of weakly noisy quantum states. Our algorithm is proved to be optimal in terms of sample complexity accompanied with polynomial classical processing time. Our result builds a bridge between the learning algorithm and quantum state complexity, meanwhile highlighting the power of learning algorithm in characterizing intrinsic properties of quantum states.

Deep learning is a powerful technique for data-driven learning in the era of Big Data. However, most deep learning models are deterministic models that ignore the uncertainty of data. Fuzzy neural networks are proposed to tackle this type of problem. In this article, we proposed a novel quantum assisted hierarchical fuzzy neural network (QA-HFNN). Different from classical fuzzy neural networks, QA-HFNN uses quantum neural networks (QNNs) to learn fuzzy membership functions. The model is a multifeature fusion learning algorithm with a parallel structural design that integrates quantum and classical neural networks. The classical network is used to capture high-dimensional neural features, the QNNs are designed to capture fuzzy logic features of the data, then, the two features are fused to form the final features to be classified. The experiment is performed on a classical computer, and the quantum circuit is built through a simulated quantum environment. The results indicate that the accuracy of QA-HFNN can equal to or even surpass classical methods in image classification tasks. The quantum circuit utilizes only a single qubit which is easy to implement. In addition, the fidelity of quantum circuit in a quantum noise environment is assessed, demonstrating that QA-HFNN has strong robustness. The time and computational complexity of QNNs was analyzed, further proving the effectiveness of the model.

Abstract Traffic classification is widely used in network security and network management. Early studies have mainly focused on mapping network traffic to different unencrypted applications, but little research has been done on network traffic classification of encrypted applications, especially the underlying traffic of encrypted applications. To address the above issues, this paper proposes a network encryption traffic classification model that combines attention mechanisms and spatiotemporal features. The model firstly uses the long short-term memory (LSTM) method to analyze continuous network flows and find the temporal correlation features between these network flows. Secondly, the convolutional neural network (CNN) method is used to extract the high-order spatial features of the network flow, and then, the squeeze and excitation (SE) module is used to weight and redistribute the high-order spatial features to obtain the key spatial features of the network flow. Finally, through the above three stages of training and learning, fast classification of network flows is achieved. The main advantages of this model are as follows: (1) the mapping relationship between network flow and label is automatically constructed by the model without manual intervention and decision by network features, (2) it has strong generalization ability and can quickly adapt to different network traffic datasets, and (3) it can handle encrypted applications and their underlying traffic with high accuracy. The experimental results show that the model can be applied to classify network traffic of encrypted and unencrypted applications at the same time, especially the classification accuracy of the underlying traffic of encrypted applications is improved. In most cases, the accuracy generally exceeds 90%.