Papers with supervised classification
Glocal: Incorporating Global Information in Local Convolution for Keyphrase Extraction (N19-1)
Copied to clipboard
| Challenge: | Graph Convolutional Networks (GCNs) model nodes’ local pairwise importance but lack the capability to model global relative importance in tasks where global ranking is a key component for the task. |
| Approach: | They propose to incorporate global relative importance information into the GCN family of models by using scaled node weights. |
| Outcome: | The proposed method improves keyphrase extraction by 2% and improves the baseline by 5%. |
MedCLIP: Contrastive Learning from Unpaired Medical Images and Text (2022.emnlp-main)
Copied to clipboard
| Challenge: | Existing vision-text contrastive learning methods encounter many false negatives, i.e., images and reports from separate patients probably carry the same semantics but are wrongly treated as negatives. |
| Approach: | They propose to decouple medical image-text contrastive learning and replace it with semantic matching loss based on medical knowledge to eliminate false negatives in contrastive training. |
| Outcome: | The proposed framework outperforms state-of-the-art methods on zero-shot prediction, supervised classification, and image-text retrieval with only 20K pre-training data. |
Data Augmentation using Machine Translation for Fake News Detection in the Urdu Language (2020.lrec-1)
Copied to clipboard
| Challenge: | supervised machine learning requires substantial amount of annotated data. |
| Approach: | They propose to use machine translation to augment annotated corpora for fake news detection in Urdu . they train a fake news classifier on an annotation dataset originally in Uru . |
| Outcome: | The proposed method fails to improve fake news detection in Urdu at the current state of machine translation quality. |
GerCCT: An Annotated Corpus for Mining Arguments in German Tweets on Climate Change (2022.lrec-1)
Copied to clipboard
| Challenge: | Recent work on annotated resources focused on single argument components, i.e., claim or evidence. |
| Approach: | They propose to annotate a German climate change argument corpus using sarcasm and toxic language to facilitate filtering out non-argumentative content. |
| Outcome: | The proposed corpus is the first to be annotated for argumentation, sarcasm and toxic language. |
An information theoretic view on selecting linguistic probes (2020.emnlp-main)
Copied to clipboard
| Challenge: | Recent advances in NLP tasks require a question of how much linguistic knowledge is encoded in neural networks. |
| Approach: | They propose to use diagnostic classifiers to perform supervised classification from internal representations. |
| Outcome: | Empirically, the two proposed criteria lead to results that agree with each other. |