Lifelong learning, also known as continual or incremental learning, is a crucial component for advancing Artificial General Intelligence (AGI) by enabling systems to continuously adapt in dynamic environments. While large language models (LLMs) have demonstrated impressive capabilities in natural language processing, existing LLM agents are typically designed for static systems and lack the ability to adapt over time in response to new challenges. This survey is the first to systematically summarize the potential techniques for incorporating lifelong learning into LLM-based agents. We categorize the core components of these agents into three modules: the perception module for multimodal input integration, the memory module for storing and retrieving evolving knowledge, and the action module for grounded interactions with the dynamic environment. We highlight how these pillars collectively enable continuous adaptation, mitigate catastrophic forgetting, and improve long-term performance. This survey provides a roadmap for researchers and practitioners working to develop lifelong learning capabilities in LLM agents, offering insights into emerging trends, evaluation metrics, and application scenarios.

The success of large language models (LLMs) has attracted many individuals to fine-tune them for domain-specific tasks by uploading their data.However, in sensitive areas like healthcare and finance, privacy concerns often arise.One promising solution is to generate synthetic data with Differential Privacy (DP) guarantees to replace private data.However, these synthetic data contain significant flawed data, which are considered as noise.Existing solutions typically rely on naive filtering by comparing ROUGE-L scores or embedding similarities, which are ineffective in addressing the noise.To address this issue, we propose RewardDS, a novel privacy-preserving framework that fine-tunes a reward proxy model and uses reward signals to guide the synthetic data generation.Our RewardDS introduces two key modules, Reward Guided Filtering and Self-Optimizing Refinement, to both filter and refine the synthetic data, effectively mitigating the noise.Extensive experiments across medical, financial, and code generation domains demonstrate the effectiveness of our method.

Tool wear prediction is of critical importance to maintain the desired part quality and improve productivity. Inspired by the successful application of deep learning in many condition monitoring tasks. In this article, a novel modeling framework is presented, which includes multiple stacked sparse autoencoders and a nonlinear regression function for tool wear prediction. Multiple stacked sparse autoencoders consists of two main structures. One model is designed with multidimensional stacked sparse autoencoders, which can learn more features from different feature domains in the raw vibration signal, and another single-dimensional stacked sparse autoencoders is used for feature fusion and deeper features learning. And a modified loss function is applied that improves the learning ability. In addition, due to the good properties of tool wear process in nonstationarity and complex nonlinear, a nonlinear regression function is utilized to enhance the progressive tool wear prediction tasks. A dataset from a real manufacturing process is used to evaluate the performance of the proposed modeling framework. Experimental results show that tool wear can be predicted accurately and stably by the proposed tool wear predictive model, which outperforms the already developed methods.