평범한 필기장

Paper : https://arxiv.org/abs/2410.02355 AlphaEdit: Null-Space Constrained Knowledge Editing for Language ModelsLarge language models (LLMs) often exhibit hallucinations due to incorrect or outdated knowledge. Hence, model editing methods have emerged to enable targeted knowledge updates. To achieve this, a prevailing paradigm is the locating-then-editing approach,arxiv.orgAbstractModel Editing ..

Paper : https://arxiv.org/abs/2509.17786 Accurate and Efficient Low-Rank Model Merging in Core SpaceIn this paper, we address the challenges associated with merging low-rank adaptations of large neural networks. With the rise of parameter-efficient adaptation techniques, such as Low-Rank Adaptation (LoRA), model fine-tuning has become more accessible. Wharxiv.orgAbstract 본 논문은 LoRA들을 common alig..

스터디 내에서 Vision Transformer를 공부하게 되었고 더 나아가 멀티모달 모델들에 대한 공부도 진행하게 되었다. 스터디에서 CoCa 논문을 발표하게 되었다. 그래서 발표 준비를 할 겸 블로그에 포스팅하게 되었다.https://arxiv.org/abs/2205.01917 CoCa: Contrastive Captioners are Image-Text Foundation ModelsExploring large-scale pretrained foundation models is of significant interest in computer vision because these models can be quickly transferred to many downstream tasks. This pap..

Transformer 자체는 이 전에 CS231n으로 대충 공부는 해봤지만 그래도 중요한 논문이기에 논문 자체를 읽어봐야겠다는 생각을 했었다. 이번 방학 때 시간이 되어 논문을 읽어보고 블로그에도 정리해보는 시간을 가졌다. 이번 논문 리뷰는 나동빈님의 논문 리뷰영상과 자료를 많이 참고해서 작성했다.https://arxiv.org/abs/1706.03762 Attention Is All You NeedThe dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration. The best performing models als..