paper-with-me

홈 › Papers

Overwriting Pretrained Bias with Finetuning Data

2023-03-10 · ICCV 2023 1 · Angelina Wang, Olga Russakovsky

Transfer learning is beneficial by allowing the expressive features of models pretrained on large-scale datasets to be finetuned for the target task of smaller, more domain-specific datasets. However, there is a concern that these pretrained models may come with their own biases which would propagate into the finetuned model. In this work, we investigate bias when conceptualized as both spurious correlations between the target task and a sensitive attribute as well as underrepresentation of a particular group in the dataset. Under both notions of bias, we find that (1) models finetuned on top of pretrained models can indeed inherit their biases, but (2) this bias can be corrected for through relatively minor interventions to the finetuning dataset, and often with a negligible impact to performance. Our findings imply that careful curation of the finetuning dataset is important for reducing biases on a downstream task, and doing so can even compensate for bias in the pretrained model.

📄 PDF Abstract BibTeX arXiv:2303.06167

Code (1)

princetonvisualai/overcoming-pretraining-bias 공식 구현 pytorch

Tasks

AttributeTransfer Learning

Similar Papers 제목 키워드 기반

Forgetting in Language Models: Capacity, Optimization, and Self-Generated Replay

2026-05-25 · Martin Marek, Dongkyu Cho, Shikai Qiu, Rumi Chunara 외 arxiv

Models trained on a new task typically degrade on prior tasks, a phenomenon known as forgetting. Traditionally, mitigating forgetting has required replaying stored exemplars from prior tasks, which is often impractical. …

Variation of Gender Biases in Visual Recognition Models Before and After Finetuning

2023-03-14 · Jaspreet Ranjit, Tianlu Wang, Baishakhi Ray, Vicente Ordonez

We introduce a framework to measure how biases change before and after fine-tuning a large scale visual recognition model for a downstream task. Deep learning models trained on increasing amounts of data are known to enc…

Object Recognition

Make Lead Bias in Your Favor: A Simple and Effective Method for News Summarization

2019-09-25 · Chenguang Zhu, ZiYi Yang, Robert Gmyr, Michael Zeng 외

Lead bias is a common phenomenon in news summarization, where early parts of an article often contain the most salient information. While many algorithms exploit this fact in summary generation, it has a detrimental effe…

News Summarization

Planted in Pretraining, Swayed by Finetuning: A Case Study on the Origins of Cognitive Biases in LLMs

2025-07-09 · Itay Itzhak, Yonatan Belinkov, Gabriel Stanovsky arxiv

Large language models (LLMs) exhibit cognitive biases -- systematic tendencies of irrational decision-making, similar to those seen in humans. Prior work has found that these biases vary across models and can be amplifie…

Analysing Moral Bias in Finetuned LLMs through Mechanistic Interpretability

2025-10-14 · Bianca Raimondi, Daniela Dalbagno, Maurizio Gabbrielli arxiv

Large language models (LLMs) have been shown to internalize human-like biases during finetuning, yet the mechanisms by which these biases manifest remain unclear. In this work, we investigated whether the well-known Knob…