The widespread adoption of Batch Normalization (BN) in contemporary deep neural architectures has demonstrated significant efficacy, particularly in the domain of Unsupervised Domain Adaptation (UDA) for cross-domain applications. Notwithstanding its success, extant BN variants often conflate source and target domain information within identical channels, potentially compromising transferability due to inter-domain feature misalignment. To address this limitation, we introduce Refined Batch Normalization (RBN), a novel normalization paradigm that leverages estimated shift to quantify discrepancies between estimated population statistics and their expected values. Our pivotal observation reveals that estimated shift can accumulate through BN stacking within the network, potentially degrading target domain performance. We elucidate how RBN mitigates this accumulation, thereby enhancing overall system efficacy. The practical implementation of this technique is realized through the RBNBlock, which supplants conventional BN with RBN in the bottleneck architecture of residual networks. Extensive empirical evaluation across diverse cross-domain benchmarks corroborates the superiority of RBN in augmenting inter-domain transferability. This perspective transcends immediate performance metrics, offering a foundational lens through which subsequent research can more deeply understand and refine the interplay between normalization strategies and domain adaptation.

In the field of E-commerce, the rapid introduction of new products poses challenges for product description generation. Traditional approaches rely on large labelled datasets, which are often unavailable for novel products with limited data. To address this issue, we propose a calibration learning approach for few-shot novel product description. Our method leverages a small amount of labelled data for calibration and utilizes the novel product's semantic representation as prompts to generate accurate and informative descriptions. We evaluate our approach on three large-scale e-commerce datasets of novel products and demonstrate its effectiveness in significantly improving the quality of generated product descriptions compared to existing methods, especially when only limited data is available. We also conduct the analysis to understand the impact of different modules on the performance.