Robustness Transfer

Robustness transfer focuses on improving the performance of machine learning models in new, unseen environments or under various perturbations by leveraging knowledge from models trained on different, potentially easier, datasets or conditions. Current research explores diverse approaches, including language-guided transfer, knowledge distillation techniques (like MixACM), and the development of novel transferability metrics (such as F-OTCE and JC-OTCE) to guide the process. This field is crucial for building more reliable and generalizable AI systems, impacting areas like computer vision, natural language processing, and robotics by enabling efficient adaptation to real-world complexities and reducing the need for extensive retraining.

Papers

March 8, 2024

Tell, Don't Show!: Language Guidance Eases Transfer Across Domains in Images and Videos
Tarun Kalluri, Bodhisattwa Prasad Majumder, Manmohan Chandraker
Language Model Gameplay Video Cross Lingual Transfer Domain Name Discriminative Information Text Supervision Robustness Transfer

October 31, 2023

Is Robustness Transferable across Languages in Multilingual Neural Machine Translation?
Leiyu Pan, Supryadi, Deyi Xiong
Native Robustness Machine Translation Unknown Language Multilingual Machine Translation Multilingual Neural Machine Translation Robustness Transfer

October 26, 2023

Fantastic Gains and Where to Find Them: On the Existence and Prospect of General Knowledge Transfer between Any Pretrained Model
Karsten Roth, Lukas Thede, Almut Sophia Koepke, Oriol Vinyals, Olivier Hénaff, Zeynep Akata
Deep Network Knowledge Transfer Visionary ProSpect Big Gain Pretrained Model Robustness Transfer

October 6, 2023

Robust Transfer Learning with Unreliable Source Data
Jianqing Fan, Cheng Gao, Jason M. Klusowski
Transfer Learning Bayes Classifier Unreliable Source Weak Signal Robustness Transfer Robust Transfer Learning

February 23, 2023

On the Hardness of Robustness Transfer: A Perspective from Rademacher Complexity over Symmetric Difference Hypothesis Space
Yuyang Deng, Nidham Gazagnadou, Junyuan Hong, Mehrdad Mahdavi, Lingjuan Lyu
Visual Perspective Hardness Result Cross Domain Generalization Rademacher Complexity Hypothesis Space Robustness Transfer Robust Domain Adaptation

February 8, 2023

Guided Learning from Demonstration for Robust Transferability
Fouad Sukkar, Victor Hernandez Moreno, Teresa Vidal-Calleja, Jochen Deuse
Robotic Manipulator Textual Demonstration Motion Planning Framework Robustness Transfer Demonstration Based Learning Accurate Motion

October 25, 2022

Accelerating Certified Robustness Training via Knowledge Transfer
Pratik Vaishnavi, Kevin Eykholt, Amir Rahmati
Adversarial Attack Native Robustness Knowledge Transfer Robust Training Robustness Transfer Training Robustness

July 12, 2022

Transferability-Guided Cross-Domain Cross-Task Transfer Learning
Yang Tan, Enming Zhang, Yang Li, Shao-Lun Huang, Xiao-Ping Zhang
Transfer Learning Transferable Representation Improving Transferability Robustness Transfer Transferability Metric

June 17, 2022

Thompson Sampling for Robust Transfer in Multi-Task Bandits
Zhi Wang, Chicheng Zhang, Kamalika Chaudhuri
Multi Task Multi Armed Bandit Thompson Sampling Robustness Transfer

June 9, 2022

On Hypothesis Transfer Learning of Functional Linear Models
Haotian Lin, Matthew Reimherr
Transfer Learning High Dimensional Reproducing Kernel Hilbert Space Robustness Transfer

November 9, 2021

MixACM: Mixup-Based Robustness Transfer via Distillation of Activated Channel Maps
Muhammad Awais, Fengwei Zhou, Chuanlong Xie, Jiawei Li, Sung-Ho Bae, Zhenguo Li
Adversarial Example Adversarial Training Adversarial Perturbation Mutual Distillation Mixup Augmentation Robustness Transfer