- 전체보기(31)
- NLP(10)
- EMNLP(7)
- ACL(6)
- cs224n(5)
- NEURIPS(4)
- 2021(3)
- GPT(3)
- Generation(3)
- 2022(3)
- Multimodal(2)
- summarization(2)
- Data-centric(2)
- Contrastive Learning(2)
- 2023(2)
- Cross-Lingual(2)
- multilingual(2)
- NAACL(2)
- chatGPT(1)
- Q&A(1)
- AAAI(1)
- simCSE(1)
- T5(1)
- keybert(1)
- Sentiment Aanlysis(1)
- dropout(1)
- LSA(1)
- BART(1)
- tokenizer(1)
- BERT(1)
- CS224(1)
- 2020(1)
- Machin Translation(1)
- Reasoning(1)
- GPT2(1)
- Microsoft(1)
- label errors(1)
- www(1)
- data(1)
- VISION(1)
- hallucination(1)
- Knowledge(1)
- IR(1)
- augmentation(1)
- Embedding(1)
- 2019(1)
- Topic modeling(1)
- word embedding(1)
- GloVe(1)
- Mutual Information(1)
- google(1)
- Zero-shot(1)
- translation(1)
- subword(1)
- Seq2Seq(1)
- OOD(1)
- IAP(1)
- knowledge base(1)
- mT6(1)
- LDA(1)
- BPE(1)
[Data-Centric AI #5] Class Imbalance, Outliers, and Distribution Shift
Introduction Data-Centric AI vs. Model-Centric AILabel ErrorsDataset Creation and CurationData-centric Evaluation of ML ModelsClass Imbalance, Outlier
[Data-Centric AI #4] Data-centric Evaluation of ML Models
Introduction Data-Centric AI vs. Model-Centric AILabel ErrorsDataset Creation and CurationData-centric Evaluation of ML ModelsClass Imbalance, Outlier
[Data-Centric AI #3] Dataset Creation and Curation
Introduction Data-Centric AI vs. Model-Centric AILabel ErrorsDataset Creation and CurationData-centric Evaluation of ML ModelsClass Imbalance, Outlier
[NLP #9] On Transferability of Prompt Tuning for Natural Language Processing (NAACL, 2022)
paper: https://aclanthology.org/2022.naacl-main.290/ > velog: https://velog.io/@zvezda/On-Transferability-of-Prompt-Tuning-for-Natural-Language-Proces
[Lab #4] DDIM: Denoising Diffusion Implicit Models (ICLR, 2021)
paper: https://arxiv.org/pdf/2010.02502.pdfbackgroundgenerative modeldiffusion model1) 이미지가 주어지면 노이지를 줄여가는 diffusion process2) noise 가지고 생성하는 sam
[NLP #8] How Far are We from Robust Long Abstractive Summarization? (EMNLP, 2022)
goal: abstractive summarization system의 수행결과를 human annotation 결과랑 비교해서 long document에 대해 얼마나 잘했는지 평가해본다. result: 비교해보니까 ROUGE result에서는 굿이었음. 그러니까 re
[NLP #7] Types of Out-of-Distribution Texts and How to Detect Them (EMNLP, 2021)
background: OOD를 탐지하는 것은 중요한 문제인데, 아직 어떠한 방식이 detect하는 데 최적의 방법인지에 대한 동의 없음. approach: OOD를 1) background shift, 2) semantic shift 둘로 나눠서 이를 탐지할 수 있는
[Lab #3-2] Cross-lingual learning for text processing: A survey (Expert Systems with Applications, 2021)
8\. Conclusions
[Lab #3-1] When is BERT Multilingual? Isolating Crucial Ingredients for Cross-lingual Transfer (NAACL, 2022)
InversionPermutationTransliterationSyntaxBilingual ModelMonolingual Model
[NLP #6] How to Adapt Your Pretrained Multilingual Model to 1600 Languages (ACL, 2021)
Goal: 1600가지 언어 데이터셋을 활용해서 기존의 PMM 모델들 performance 검증Challenge: 1) 소량 2) narrow domainResult: XLM-R good
[Lab #2] Training language models to follow instructions with human feedback (NeurIPS, 2022)
paper: https://proceedings.neurips.cc/paper_files/paper/2022/file/b1efde53be364a73914f58805a001731-Paper-Conference.pdfBackground: LM 모델을 더 크게 만든
[Lab #1] Efficient Dialogue State Tracking by Selectively Overwriting Memory (ACL, 2020)
Dialogue systemopen-domain dialogue system : 대화주제 자유롭게task oriented dialogue (TOD) system: 특정 task 수행TODsingle domain TOD: 하나의 작업만 수행Multi domain TOD:
[Data-Centric AI #2] Label Errors (Confident Learning)
Introduction Data-Centric AI vs. Model-Centric AILabel ErrorsDataset Creation and CurationData-centric Evaluation of ML ModelsClass Imbalance, Outlier
[Data-Centric AI #1] Data-Centric AI vs. Model-Centric AI
Introduction Data-Centric AI vs. Model-Centric AILabel ErrorsDataset Creation and CurationData-centric Evaluation of ML ModelsClass Imbalance, Outlier
[Data-Centric AI #0] Introduction (IAP, 2023)
page: https://dcai.csail.mit.edu/1/17/23: Data-Centric AI vs. Model-Centric AI1/18/23: Label Errors1/19/23: Dataset Creation and Curation1/20/23:
[NLP #5] Mutual Information Alleviates Hallucinations in Abstractive Summarization (EMNLP, 2022)
paper: https://aclanthology.org/2022.emnlp-main.399.pdfcode: https://github.com/VanderpoelLiam/CPMIBackground: hallucinate: 요약한게 source docu
[NLP #4] BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension (ACL, 2020)
Paper: https://aclanthology.org/2020.acl-main.703/ > Code: > - Ko-BART : https://github.com/SKT-AI/KoBART
[NLP #3] SimCSE: Simple Contrastive Learning of Sentence Embeddings (EMNLP, 2021)
한줄 요약: unlabeled or labeled data 모두에서 setence embedding 뽑을 수 있다? Paper: https://aclanthology.org/2021.emnlp-main.552/Code: https://github.co
[Multimodal #2] Found in Translation: Learning Robust Joint Representations by Cyclic Translations between Modalities (AAAI, 19)
Paper: https://ojs.aaai.org/index.php/AAAI/article/view/4666 > Github: https://github.com/hainow/MCTN # Introduction # Related Work # Proposed Appro
[Multimodal #1] ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Language Tasks (NeurIPS, 2019)
인용이 무려 1913회! ㅇ_ㅇ Summary Introduction Approach Preliminaries: Bidirectional Encoder Representations from Transformers (BERT) Text Representation Tr