Existing adaptive learning systems struggle to simultaneously achieve deep personalization, dynamic adaptability, and content trustworthiness, particularly in logically rigorous STEM fields where Large Language Models (LLMs) are prone to "hallucination".This paper introduces LEARNERCOMPASS (Cognitive Multimodel Planning & Adaptive System), an integrated, end-to-end framework for adaptive learning.At its core, the framework features a novel multi-model path planning algorithm that orchestrates and fuses the outputs of heterogeneous LLM experts to generate and optimize learning sequences.To enable deep personalization, we design a dynamic cognitive diagnosis module that employs an innovative encoderdecoder architecture to generate precise, multidimensional cognitive state vectors for learners.To ensure trustworthiness, the system leverages an adaptively constructed dynamic knowledge graph and a Graph-RAG mechanism to provide factual anchors and logical constraints for LLM reasoning, thereby mitigating hallucinations.Extensive experiments demonstrate that LEARNERCOMPASS significantly outperforms state-of-the-art baselines in generating high-quality personalized learning paths.Furthermore, ablation studies validate the critical contributions of our dynamic cognitive diagnosis and multi-model planning components.

Xueyang Zhou, Weidong Wang, Lin Lu, Jiawen Shi, Guiyao Tie, Xu Yongtian, Lixing Chen, Pan Zhou, Neil Zhenqiang Gong, Lichao Sun. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

In recent years, the frequent leakage of sensitive information in natural language texts has presented significant challenges in text distribution, sharing, and migration. Existing methods for text protection and watermarking are limited, capable of either hide sensitive data or prevent unauthorized copying and tampering, but not both simultaneously. To address these limitations, this paper proposes a masking-watermarking cooperative framework designed to hide text-sensitive information, prevent unintentional data leakage, and ensure content ownership verification and tampering prevention. The framework introduces three novel techniques: a variable autoencoder to ensure diverse watermark generation, an improved transformer to enhance the adaptability of dynamic masks, and a dual discriminator for joint verification of text and watermarks. A comprehensive evaluation was conducted, covering text similarity, masking flexibility, and the imperceptibility of text masking, as well as watermark classification recognition and robustness against various attacks. The proposed framework achieved a score of 0.98 on the SBERT metric, demonstrating its effectiveness in achieving imperceptible text masking and robust watermark embedding.

Self-correction of large language models (LLMs) emerges as a critical component for enhancing their reasoning performance. Although various self-correction methods have been proposed, a comprehensive evaluation of these methods remains largely unexplored, and the question of whether LLMs can truly correct themselves is a matter of significant interest and concern. In this study, we introduce CorrectBench, a benchmark developed to evaluate the effectiveness of self-correction strategies, including intrinsic, external, and fine-tuned approaches, across three tasks: commonsense reasoning, mathematical reasoning, and code generation. Our findings reveal that: 1) Self-correction methods can improve accuracy, especially for complex reasoning tasks; 2) Mixing different self-correction strategies yields further improvements, though it reduces efficiency; 3) Reasoning LLMs (e.g., DeepSeek-V3) have limited optimization under additional self-correction methods and have high time costs. Interestingly, a comparatively simple chain-of-thought (CoT) baseline demonstrates competitive accuracy and efficiency. These results underscore the potential of self-correction to enhance LLM's reasoning performance while highlighting the ongoing challenge of improving their efficiency. Consequently, we advocate for further research focused on optimizing the balance between reasoning capabilities and operational efficiency.

Vision-Language-Action (VLA) models have advanced robotic control by enabling end-to-end decision-making directly from multimodal inputs. However, their tightly coupled architectures expose novel security vulnerabilities. Unlike traditional adversarial perturbations, backdoor attacks represent a stealthier, persistent, and practically significant threat—particularly under the emerging Training-as-a-Service paradigm—but remain largely unexplored in the context of VLA models. To address this gap, we propose **BadVLA**, a backdoor attack method based on Objective-Decoupled Optimization, which for the first time exposes the backdoor vulnerabilities of VLA models. Specifically, it consists of a two-stage process: (1) explicit feature-space separation to isolate trigger representations from benign inputs, and (2) conditional control deviations that activate only in the presence of the trigger, while preserving clean-task performance. Empirical results on multiple VLA benchmarks demonstrate that BadVLA consistently achieves near-100\% attack success rates with minimal impact on clean task accuracy. Further analyses confirm its robustness against common input perturbations, task transfers, and model fine-tuning, underscoring critical security vulnerabilities in current VLA deployments. Our work offers the first systematic investigation of backdoor vulnerabilities in VLA models, highlighting an urgent need for secure and trustworthy embodied model design practices.