Unsupervised optimal deep transfer learning for classification under general conditional shift
Holder： Yukun Liu（East China Normal University)
Time：2025-03-11 14:00-15:00
Location：Siyuan Lecture Hall，Zhi Hua Building-225
Abstract:
Classifiers trained solely on labeled source data may yield misleading results when applied to unlabeled target data drawn from a different distribution. Transfer learning can rectify this by transferring knowledge from source to target data, but its effectiveness frequently relies on stringent assumptions, such as label shift. In this paper, we introduce a novel General Conditional Shift (GCS) assumption, which encompasses label shift as a special scenario. Under GCS, we demonstrate that both the target distribution and the shift function are identifiable under mild conditions. To estimate the conditional probabilities ${\bm\eta}_P$
for source data, we propose leveraging deep neural networks (DNNs). Subsequent to transferring the DNN estimator, we estimate the target label distribution ${\bm\pi}_Q$ utilizing a pseudo-maximum likelihood approach. Ultimately, by incorporating these estimates and circumventing the need to estimate the shift function, we construct our proposed Bayes classifier. We establish concentration bounds for our estimators of both {\bm\eta}_P$ and ${\bm\pi}_Q$ in terms of the intrinsic dimension of ${\bm\eta}_P$. Notably, our DNN-based classifier achieves the optimal minimax rate, up to a logarithmic factor. A key advantage of our method is its capacity to effectively combat the curse of dimensionality when ${\bm\eta}_P$ exhibits a low-dimensional structure. Numerical simulations, along with an analysis of an Alzheimer's disease dataset, underscore its exceptional performance.
About the Speaker:
刘玉坤，华东师范大学统计学院教授。本科和博士毕业于南开大学统计系，之后一直在华东师范大学任教。研究兴趣包括分布偏移数据、因果推断、共形推断、半参数统计和机器学习等。成果曾发表在JRSSB、AOS、JASA、Biometrika、JOE等期刊; 主持科技部国家重点研发计划课题和国家自然科学基金项目；入选国家高层次青年人才计划。
Your participation is warmly welcomed!
欢迎扫码关注北大统计科学中心公众号，了解更多讲座信息!
