Readings
Natural Language Processing (NLP)
Language Modeling
Paper Title | Year | Reason |
---|---|---|
Hopfield Networks is All You Need | 2020 | Looks at the Transformer and BERT architecture from the perspective of Hopfield networks. |
Pre-training via Paraphrasing | 2020 | An improvement over the Masked Language Modeling techniques. |
Language Models are Few-Shot Learners | 2020 | The famous GPT3 model. Needs no other reason. |
Big Bird - Transformers for Longer Sequences | 2020 | Sparse attention mechanism enabling processing of longer sequences |
LayoutLM: Pre-training of Text and Layout for Document Image Understanding | 2020 | Incorporating layout (position and image) embeddings to BERT pre-training |
Exploring the Limits of Language Modeling | 2016 | |
Don’t Stop Pretraining: Adapt Language Models to Domains and Tasks | 2020 | Investigates the effectiveness of using task-related data in pretraining LMs |
Controllable Text Generation
Paper Title | Year | Reason |
---|---|---|
Exploring Controllable Text Generation Techniques | 2020 | Great work of putting controllable LG papers into a unified framework. |
Machine Translation (MT)
Paper Title | Year | Reason |
---|---|---|
Convolutional Sequence to Sequence Learning | 2017 | Introduced convolution operation to Seq2Seq problems |
Relation Extraction (RE)
Paper Title | Year | Reason |
---|---|---|
Simple BERT Models for Relation Extraction and Semantic Role Labeling | 2019 | Applying BERT to relation extraction |
Semantic Role Labeling (SRL)
Paper Title | Year | Reason |
---|---|---|
Simple BERT Models for Relation Extraction and Semantic Role Labeling | 2019 | Applying BERT to semantic role labeling |
Open Information Extraction (OpenIE)
Paper Title | Year | Reason |
---|---|---|
Supervised Open Information Extraction | 2018 | Formulation of framing OpenIE as a sequence tagging problem |
Document Intelligence
Paper Title | Year | Reason |
---|---|---|
LayoutLM: Pre-training of Text and Layout for Document Image Understanding | 2020 | Incorporating layout (position and image) embeddings to BERT pre-training |
Computer Vision (CV)
Other
Paper Title | Year | Reason |
---|---|---|
Towards Learning Convolutions from Scratch | 2020 |
Time Series Analysis
Paper Title | Year | Reason |
---|---|---|
Temporal Fusion Transformers for Interpretable Multi-horizon Time Series Forecasting | 2019 |