Recent advances in neural decoding have enabled the reconstruction of visual experiences from brain activity, positioning fMRI-to-image reconstruction as a promising bridge between neuroscience and computer vision. However, current methods predominantly rely on subject-specific models or require subject-specific fine-tuning, limiting their scalability and real-world applicability. In this work, we introduce ZEBRA, the first zero-shot brain visual decoding framework that eliminates the need for subject-specific adaptation. ZEBRA is built on the key insight that fMRI representations can be decomposed into subject-related and semantic-related components. By leveraging adversarial training, our method explicitly disentangles these components to isolate subject-invariant, semantic-specific representations. This disentanglement allows ZEBRA to generalize to unseen subjects without any additional fMRI data or retraining. Extensive experiments show that ZEBRA significantly outperforms zero-shot baselines and achieves performance comparable to fully finetuned models on several metrics. Our work represents a scalable and practical step toward universal neural decoding. Code and model weights are available at: https://github.com/xmed-lab/ZEBRA.

Current Large Language Models (LLMs) have shown strong reasoning capabilities in commonsense question answering benchmarks, but the process underlying their success remains largely opaque. As a consequence, recent approaches have equipped LLMs with mechanisms for knowledge retrieval, reasoning and introspection, not only to improve their capabilities but also to enhance the interpretability of their outputs. However, these methods require additional training, hand-crafted templates or human-written explanations. To address these issues, we introduce ZEBRA, a zero-shot question answering framework that combines retrieval, case-based reasoning and introspection and dispenses with the need for additional training of the LLM. Given an input question, ZEBRA retrieves relevant question-knowledge pairs from a knowledge base and generates new knowledge by reasoning over the relationships in these pairs. This generated knowledge is then used to answer the input question, improving the model's performance and interpretability. We evaluate our approach across 8 well-established commonsense reasoning benchmarks, demonstrating that ZEBRA consistently outperforms strong LLMs and previous knowledge integration approaches, achieving an average accuracy improvement of up to 4.5 points.

Recently, deep learning model has received substantial interest in industrial applications [4, 5]. In the vision-based An automatic vision-based sewer inspection plays a key sewer inspection community, deep learning also attracts increasing role of sewage system in a modern city. Recent advances focus attention from both academia and industry [6, 7, 8]. on utilizing deep learning model to realize the sewer inspection Here, we focus on the sewer defect classification in the setting system, benefiting from the capability of data-driven of multi-label, in which multiply defect classes in an feature representation. However, the inherent uncertainty of image are recognized simultaneously. Although these deep sewer defects is ignored, resulting in the missed detection learning-based methods have achieved acceptable performances of serious unknown sewer defect categories. In this paper, of sewer defect classification, while the inherent uncertainty we propose a trustworthy multi-label sewer defect classification of sewer defects might not be considered sufficiently (TMSDC) method, which can quantify the uncertainty of in real-world applications [9].