The field of machine learning is witnessing significant developments in classification and knowledge distillation. Researchers are focusing on improving the performance of extreme classification tasks, particularly for infrequent categories, by leveraging robust architectures and knowledge transfer. Another area of interest is the application of deep learning-based risk models for breast cancer detection, where accurate spatial alignment of longitudinal mammograms is crucial for personalized screening. Moreover, there is a growing trend towards exploring cross-modal knowledge distillation, which enables the transfer of knowledge between different modalities with limited semantic overlap. Noteworthy papers in this area include: The paper on LEVER, which presents a novel solution for addressing the challenges posed by underperforming infrequent categories in extreme classification tasks. The study on the impact of longitudinal mammogram alignment on breast cancer risk assessment, which highlights the importance of image-based deformation fields for spatial alignment. The work on Asymmetric Cross-modal Knowledge Distillation, which proposes a framework for bridging modalities with weak semantic consistency. The paper on Enriching Knowledge Distillation with Cross-Modal Teacher Fusion, which explores the use of CLIP's vision-language knowledge as a complementary source of supervision for knowledge distillation. The research on Revisiting Cross-Architecture Distillation, which proposes a Dual-Teacher Knowledge Distillation framework for transferring knowledge from Vision Transformers to lightweight CNNs.