공대생 도전 일지

세부 정리 - PrivacyRestore

https://aclanthology.org/2025.acl-long.532/ PrivacyRestore: Privacy-Preserving Inference in Large Language Models via Privacy Removal and RestorationZiqian Zeng, Jianwei Wang, Junyao Yang, Zhengdong Lu, Haoran Li, Huiping Zhuang, Cen Chen. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.aclanthology.orgACL 2025에 붙은 논문입니다. LLM ..

딥러닝 응용 시험 정리 -2

https://arxiv.org/abs/2309.15649 Generative Speech Recognition Error Correction with Large Language Models and Task-Activating PromptingWe explore the ability of large language models (LLMs) to act as speech recognition post-processors that perform rescoring and error correction. Our first focus is on instruction prompting to let LLMs perform these task without fine-tuning, for which we evarxiv...

딥러닝 응용 시험 정리 - 1 CTC Loss, LoRA

일단 기말고사가 닥쳐와서.... https://docs.pytorch.org/docs/stable/generated/torch.nn.CTCLoss.html CTCLoss — PyTorch 2.9 documentationCTCLoss class torch.nn.CTCLoss(blank=0, reduction='mean', zero_infinity=False)[source] The Connectionist Temporal Classification loss. Calculates loss between a continuous (unsegmented) time series and a target sequence. CTCLoss sums over the probability odocs.pytorch.org일단 ..

조금은 이제 알 것 같은....https://aclanthology.org/2025.acl-long.58/ ObfusLM: Privacy-preserving Language Model Service against Embedding Inversion AttacksYu Lin, Ruining Yang, Yunlong Mao, Qizhi Zhang, Jue Hong, Quanwei Cai, Ye Wu, Huiqi Liu, Zhiyu Chen, Bing Duan, Sheng Zhong. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.aclanthol..

https://arxiv.org/abs/2311.06805 Tunable Soft Prompts are Messengers in Federated LearningFederated learning (FL) enables multiple participants to collaboratively train machine learning models using decentralized data sources, alleviating privacy concerns that arise from directly sharing local data. However, the lack of model privacy protectionarxiv.orgTunable Soft Prompts are Messengers in Fede..

https://github.com/MyoungJinKim/AAA737_TermProject GitHub - MyoungJinKim/AAA737_TermProject: Prompting Large Language Models with Speech Recognition Abilities 논문 코드 재현Prompting Large Language Models with Speech Recognition Abilities 논문 코드 재현 - MyoungJinKim/AAA737_TermProjectgithub.com코드가 없어서 직접 구현했습니다.....README 읽어보시면 잘 아실 수 있습니다.코드는 아래 레포지에서 하나 하나 따오거나, 참조해서 만들었습니다. https://github.com/sooftware..

논문 작성을 위해 이어서...https://openreview.net/forum?id=MyRcW53CCC PrivacyRestore: Privacy-Preserving Inference in Large Language...The widespread usage of online Large Language Models (LLMs) inference services has raised significant privacy concerns about the potential exposure of private information in user inputs to...openreview.netPrivacyRestore: Privacy-Preserving Inference in Large Language Models..

https://aclanthology.org/2022.naacl-main.231/ Privacy-Preserving Text Classification on BERT Embeddings with Homomorphic EncryptionGaram Lee, Minsoo Kim, Jai Hyun Park, Seung-won Hwang, Jung Hee Cheon. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022.aclanthology.orgPrivacy-Preserving Text Classif..

https://arxiv.org/abs/2510.23274 Privacy-Preserving Semantic Communication over Wiretap Channels with Learnable Differential PrivacyWhile semantic communication (SemCom) improves transmission efficiency by focusing on task-relevant information, it also raises critical privacy concerns. Many existing secure SemCom approaches rely on restrictive or impractical assumptions, such as favoraarxiv.orgP..

https://aclanthology.org/2025.acl-long.1394/ Pretraining Context Compressor for Large Language Models with Embedding-Based MemoryYuhong Dai, Jianxun Lian, Yitian Huang, Wei Zhang, Mingyang Zhou, Mingqi Wu, Xing Xie, Hao Liao. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.aclanthology.org긴 컨텍스트 처리하는데 드는 비용과 메모리 사용량이 매우 큼! => ..