반응형 arxiv572 [2022-11-12] 오늘의 자연어처리 Exploring Robustness of Prefix Tuning in Noisy Data: A Case Study in Financial Sentiment Analysis The invention of transformer-based models such as BERT, GPT, and RoBERTa has enabled researchers and financial companies to finetune these powerful models and use them in different downstream tasks to achieve state-of-the-art performance. Recently, a lightweight alternative (approximately 0.1% - 3% .. 2022. 11. 12. [2022-11-11] 오늘의 자연어처리 Evaluating and Improving Context Attention Distribution on Multi-Turn Response Generation using Self-Contained Distractions Despite the rapid progress of open-domain generation-based conversational agents, most deployed systems treat dialogue contexts as single-turns, while systems dealing with multi-turn contexts are less studied. There is a lack of a reliable metric for evaluating multi-turn m.. 2022. 11. 11. [2022-11-10] 오늘의 자연어처리 SocioProbe: What, When, and Where Language Models Learn about Sociodemographics Pre-trained language models (PLMs) have outperformed other NLP models on a wide range of tasks. Opting for a more thorough understanding of their capabilities and inner workings, researchers have established the extend to which they capture lower-level knowledge like grammaticality, and mid-level semantic knowledge l.. 2022. 11. 10. [2022-11-09] 오늘의 자연어처리 How Much Does Attention Actually Attend? Questioning the Importance of Attention in Pretrained Transformers The attention mechanism is considered the backbone of the widely-used Transformer architecture. It contextualizes the input by computing input-specific attention matrices. We find that this mechanism, while powerful and elegant, is not as important as typically thought for pretrained langu.. 2022. 11. 9. 이전 1 ··· 109 110 111 112 113 114 115 ··· 143 다음 반응형