Researchers have developed a new AI model called CFM that can understand the concepts behind images and provide explanations for its decisions. Unlike previous models, CFM provides fine-grained, human-interpretable concepts that are spatially grounded in the input image. This means it can explain not just classification tasks but also more complex tasks like segmentation and captioning. The model is based on a foundation model with strong semantic representations, which allows for explanations of downstream tasks. By analyzing local co-occurrence dependencies of concepts, CFM can even improve concept naming and provide richer explanations. --- Why it matters: This matters to AI researchers because it provides a way to understand the decision-making process behind complex vision models, making them more interpretable and potentially leading to better performance on real-world tasks. Source: https://arxiv.org/abs/2601.13798