no code implementations • 5 Jan 2024 • Stephen Obadinma, Xiaodan Zhu, Hongyu Guo
We introduce a new framework of adversarial attacks, named calibration attacks, in which the attacks are generated and organized to trap victim models to be miscalibrated without altering their original accuracy, hence seriously endangering the trustworthiness of the models and any decision-making based on their confidence scores.
no code implementations • 5 Mar 2023 • Stephen Obadinma, Hongyu Guo, Xiaodan Zhu
In this paper, we examine the effectiveness of several popular task-agnostic data augmentation techniques, i. e., EDA, Back Translation, and Mixup, when using two general parameter efficient tuning methods, P-tuning v2 and LoRA, under data scarcity.
1 code implementation • 7 Feb 2023 • Stephen Obadinma, Faiza Khan Khattak, Shirley Wang, Tania Sidhom, Elaine Lau, Sean Robertson, Jingcheng Niu, Winnie Au, Alif Munim, Karthik Raja K. Bhaskar, Bencheng Wei, Iris Ren, Waqar Muhammad, Erin Li, Bukola Ishola, Michael Wang, Griffin Tanner, Yu-Jia Shiah, Sean X. Zhang, Kwesi P. Apponsah, Kanishk Patel, Jaswinder Narain, Deval Pandya, Xiaodan Zhu, Frank Rudzicz, Elham Dolatabadi
Building Agent Assistants that can help improve customer service support requires inputs from industry users and their customers, as well as knowledge about state-of-the-art Natural Language Processing (NLP) technology.
no code implementations • 29 Sep 2021 • Stephen Obadinma, Xiaodan Zhu, Hongyu Guo
Our studies suggest the following: most of the time curriculum learning has a negligible effect on calibration, but in certain cases under the context of limited training time and noisy data, curriculum learning can substantially reduce calibration error in a manner that cannot be explained by dynamically sampling the dataset.
1 code implementation • SEMEVAL 2020 • Xiaoyu Yang, Stephen Obadinma, Huasha Zhao, Qiong Zhang, Stan Matwin, Xiaodan Zhu
Subtask-1 aims to determine whether a given sentence is a counterfactual statement or not.