Alert button
Picture for Timo Schick

Timo Schick

Alert button

Generating Datasets with Pretrained Language Models

Apr 17, 2021
Timo Schick, Hinrich Schütze

Figure 1 for Generating Datasets with Pretrained Language Models
Figure 2 for Generating Datasets with Pretrained Language Models
Figure 3 for Generating Datasets with Pretrained Language Models
Figure 4 for Generating Datasets with Pretrained Language Models
Viaarxiv icon

Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP

Feb 28, 2021
Timo Schick, Sahana Udupa, Hinrich Schütze

Figure 1 for Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP
Figure 2 for Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP
Figure 3 for Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP
Figure 4 for Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP
Viaarxiv icon

Few-Shot Text Generation with Pattern-Exploiting Training

Dec 22, 2020
Timo Schick, Hinrich Schütze

Figure 1 for Few-Shot Text Generation with Pattern-Exploiting Training
Figure 2 for Few-Shot Text Generation with Pattern-Exploiting Training
Figure 3 for Few-Shot Text Generation with Pattern-Exploiting Training
Figure 4 for Few-Shot Text Generation with Pattern-Exploiting Training
Viaarxiv icon

Automatically Identifying Words That Can Serve as Labels for Few-Shot Text Classification

Oct 26, 2020
Timo Schick, Helmut Schmid, Hinrich Schütze

Figure 1 for Automatically Identifying Words That Can Serve as Labels for Few-Shot Text Classification
Figure 2 for Automatically Identifying Words That Can Serve as Labels for Few-Shot Text Classification
Figure 3 for Automatically Identifying Words That Can Serve as Labels for Few-Shot Text Classification
Figure 4 for Automatically Identifying Words That Can Serve as Labels for Few-Shot Text Classification
Viaarxiv icon

It's Not Just Size That Matters: Small Language Models Are Also Few-Shot Learners

Sep 15, 2020
Timo Schick, Hinrich Schütze

Figure 1 for It's Not Just Size That Matters: Small Language Models Are Also Few-Shot Learners
Figure 2 for It's Not Just Size That Matters: Small Language Models Are Also Few-Shot Learners
Figure 3 for It's Not Just Size That Matters: Small Language Models Are Also Few-Shot Learners
Figure 4 for It's Not Just Size That Matters: Small Language Models Are Also Few-Shot Learners
Viaarxiv icon

Exploiting Cloze Questions for Few-Shot Text Classification and Natural Language Inference

Jan 21, 2020
Timo Schick, Hinrich Schütze

Figure 1 for Exploiting Cloze Questions for Few-Shot Text Classification and Natural Language Inference
Figure 2 for Exploiting Cloze Questions for Few-Shot Text Classification and Natural Language Inference
Figure 3 for Exploiting Cloze Questions for Few-Shot Text Classification and Natural Language Inference
Figure 4 for Exploiting Cloze Questions for Few-Shot Text Classification and Natural Language Inference
Viaarxiv icon

BERTRAM: Improved Word Embeddings Have Big Impact on Contextualized Model Performance

Nov 07, 2019
Timo Schick, Hinrich Schütze

Figure 1 for BERTRAM: Improved Word Embeddings Have Big Impact on Contextualized Model Performance
Figure 2 for BERTRAM: Improved Word Embeddings Have Big Impact on Contextualized Model Performance
Figure 3 for BERTRAM: Improved Word Embeddings Have Big Impact on Contextualized Model Performance
Figure 4 for BERTRAM: Improved Word Embeddings Have Big Impact on Contextualized Model Performance
Viaarxiv icon

Rare Words: A Major Problem for Contextualized Embeddings And How to Fix it by Attentive Mimicking

Apr 20, 2019
Timo Schick, Hinrich Schütze

Figure 1 for Rare Words: A Major Problem for Contextualized Embeddings And How to Fix it by Attentive Mimicking
Figure 2 for Rare Words: A Major Problem for Contextualized Embeddings And How to Fix it by Attentive Mimicking
Figure 3 for Rare Words: A Major Problem for Contextualized Embeddings And How to Fix it by Attentive Mimicking
Figure 4 for Rare Words: A Major Problem for Contextualized Embeddings And How to Fix it by Attentive Mimicking
Viaarxiv icon