Characterizing and avoiding negative transfer

Z Wang, Z Dai, B Póczos… - Proceedings of the IEEE …, 2019 - openaccess.thecvf.com
When labeled data is scarce for a specific target task, transfer learning often offers an
effective solution by utilizing data from a related source task. However, when transferring …

A survey on negative transfer

W Zhang, L Deng, L Zhang, D Wu - IEEE/CAA Journal of …, 2022 - ieeexplore.ieee.org
Transfer learning (TL) utilizes data or knowledge from one or more source domains to
facilitate learning in a target domain. It is particularly useful when the target domain has very …

Gradient vaccine: Investigating and improving multi-task optimization in massively multilingual models

Z Wang, Y Tsvetkov, O Firat, Y Cao - arXiv preprint arXiv:2010.05874, 2020 - arxiv.org
Massively multilingual models subsuming tens or even hundreds of languages pose great
challenges to multi-task optimization. While it is a common practice to apply a language …

On negative interference in multilingual models: Findings and a meta-learning treatment

Z Wang, ZC Lipton, Y Tsvetkov - arXiv preprint arXiv:2010.03017, 2020 - arxiv.org
Modern multilingual models are trained on concatenated text from multiple languages in
hopes of conferring benefits to each (positive transfer), with the most pronounced benefits …

Efficient meta lifelong-learning with limited memory

Z Wang, SV Mehta, B Póczos, J Carbonell - arXiv preprint arXiv …, 2020 - arxiv.org
Current natural language processing models work well on a single task, yet they often fail to
continuously learn new tasks without forgetting previous ones as they are re-trained …

A novel active multi-source transfer learning algorithm for time series forecasting

Q Gu, Q Dai - Applied Intelligence, 2021 - Springer
Abstract In Time Series Forecasting (TSF), researchers usually assume that there is enough
training data can be obtained, with the old and new data satisfying the same distribution …

Integrating multi-source transfer learning, active learning and metric learning paradigms for time series prediction

Q Gu, Q Dai, H Yu, R Ye - Applied Soft Computing, 2021 - Elsevier
Abstract Traditional Time Series Prediction (TSP) algorithms assume that the training and
testing data follow the same distribution and a large amount of data can be obtained …

Federated graph learning for cross-domain recommendation

Z Yang, Z Peng, Z Wang, J Qi, C Chen, W Pan… - arXiv preprint arXiv …, 2024 - arxiv.org
Cross-domain recommendation (CDR) offers a promising solution to the data sparsity
problem by enabling knowledge transfer across source and target domains. However, many …

Mitigating Negative Transfer in Cross-Domain Recommendation via Knowledge Transferability Enhancement

Z Song, W Zhang, L Deng, J Zhang, Z Wu… - Proceedings of the 30th …, 2024 - dl.acm.org
Cross-Domain Recommendation (CDR) is a promising technique to alleviate data sparsity
by transferring knowledge across domains. However, the negative transfer issue in the …

A simple approach to balance task loss in multi-task learning

S Liang, C Deng, Y Zhang - … Conference on Big Data (Big Data …, 2021 - ieeexplore.ieee.org
In multi-task learning, the training losses of different tasks are varying. There are many works
to handle this situation and we classify them into five categories. In this paper, we propose a …