Differences

This shows you the differences between two versions of the page.

--- nlp:domain_adaptation [2023/03/02 04:21] – [Papers (General Domain Adaptation in NLP)] jmflanig
+++ nlp:domain_adaptation [2023/06/15 07:36] (current) – external edit 127.0.0.1
@@ Line 5: / Line 5: @@
   * [[https://arxiv.org/pdf/1812.02849.pdf|2018 - A Survey of Unsupervised Deep Domain Adaptation]] Unsupervised domain adaptation differs from regular domain adaptation in that you don't get to see labeled examples in the domain of interest, only unlabeled examples.
   * [[https://arxiv.org/pdf/1806.00258.pdf|Chu & Wang 2018 - A Survey of Domain Adaptation for Neural Machine Translation]]
+  * [[https://arxiv.org/pdf/2006.00632.pdf|Ramponi & Plank 2020 - Neural Unsupervised Domain Adaptation in NLP - A Survey]]
+===== Domain Adaptation (Outside of NLP) =====
+  * [[https://proceedings.mlr.press/v37/ganin15.pdf|Ganin & Lempitsky 2015 - Unsupervised Domain Adaptation by Backpropagation]]
+  * [[https://arxiv.org/pdf/1505.07818.pdf|Ganin et al 2016 - Domain-Adversarial Training of Neural Networks]]
 ===== Papers (General Domain Adaptation in NLP) =====
-See also [[https://www.aminer.org/topic/5faa648192c7f9be21f70f88|Awesome Neural Adaptation in NLP]] A curated list of unsupervised domain adaption papers in NLP (not including MT).
+See also [[https://github.com/bplank/awesome-neural-adaptation-in-NLP|Awesome Neural Adaptation in NLP]] or [[https://www.aminer.org/topic/5faa648192c7f9be21f70f88|here (old)]] A curated list of unsupervised domain adaption papers in NLP (not including MT).
   * [[https://arxiv.org/pdf/0907.1815.pdf|Daumé III 2009 - Frustratingly Easy Domain Adaptation]]  A seminal paper, the baseline that you should always try (for linear models).
   * The obvious baseline for neural networks is to fine-tune a pre-trained network on the new domain.  I don't know any papers looking into this method and the trade-off associated with it, but there should be one
+  * [[https://arxiv.org/pdf/1505.07818.pdf|Ganin et al 2016 - Domain-Adversarial Training of Neural Networks]]
   * [[https://www.aclweb.org/anthology/C16-1038.pdf|Kim et al 2016 - Frustratingly Easy Neural Domain Adaptation]]  Not a very good paper. Didn't compare to sensible baselines such as fine-tuning on the new domain.
   * [[https://www.aclweb.org/anthology/K17-1029.pdf|2017 - Neural Domain Adaptation for Biomedical Question Answering]]
@@ Line 18: / Line 24: @@
   * [[https://aclanthology.org/2021.adaptnlp-1.22.pdf|Xu et al 2021 - Gradual Fine-Tuning for Low-Resource Domain Adaptation]]
   * [[https://aclanthology.org/2021.emnlp-main.566.pdf|Kulshreshtha et al 2021 - Back-Training excels Self-Training at Unsupervised Domain Adaptation of Question Generation and Passage Retrieval]] Back-training: generating noisy inputs given outputs (vs self-training: generating noisy outputs given inputs).
+  * [[https://aclanthology.org/2022.naacl-main.96.pdf|Chronopoulou et al 2022 - Efficient Hierarchical Domain Adaptation for Pretrained Language Models]]
+  * [[https://arxiv.org/pdf/2302.03169.pdf|Xie et al 2023 - Data Selection for Language Models via Importance Resampling]]
+  * [[https://arxiv.org/pdf/2302.03194.pdf|Malik et al 2023 - UDApter - Efficient Domain Adaptation Using Adapters]]
+{{media:domain_adaptation_table2020.png}}\\
+Table from [[https://arxiv.org/pdf/2006.00632.pdf|Ramponi & Plank 2020]].
 ===== Domain Adaptation in NLP Tasks =====
@@ Line 27: / Line 39: @@
 ===== Related Pages =====
   * [[Data Augmentation]]
+  * [[ml:Fine-Tuning]]
   * [[ml:Semi-supervised Learning]]