中文
相关论文

相关论文: Does Vec2Text Pose a New Corpus Poisoning Threat?

200 篇论文

With the help of conditioning mechanisms, the state-of-the-art diffusion models have achieved tremendous success in guided image generation, particularly in text-to-image synthesis. To gain a better understanding of the training process and…

密码学与安全 · 计算机科学 2023-10-24 Shengfang Zhai , Yinpeng Dong , Qingni Shen , Shi Pu , Yuejian Fang , Hang Su

High-quality paraphrases are easy to produce using instruction-tuned language models or specialized paraphrasing models. Although this capability has a variety of benign applications, paraphrasing attacks$\unicode{x2013}$paraphrases applied…

计算与语言 · 计算机科学 2025-03-21 Rafael Rivera Soto , Barry Chen , Nicholas Andrews

Large language models (LLMs) have demonstrated impressive natural language processing abilities but face challenges such as hallucination and outdated knowledge. Retrieval-Augmented Generation (RAG) has emerged as a state-of-the-art…

密码学与安全 · 计算机科学 2026-01-09 Baolei Zhang , Yuxi Chen , Zhuqing Liu , Lihai Nie , Tong Li , Zheli Liu , Minghong Fang

Deep neural networks are facing a potential security threat from adversarial examples, inputs that look normal but cause an incorrect classification by the deep neural network. For example, the proposed threat could result in hand-written…

计算机视觉与模式识别 · 计算机科学 2016-10-17 Abigail Graese , Andras Rozsa , Terrance E. Boult

Textual data is often represented as real-numbered embeddings in NLP, particularly with the popularity of large language models (LLMs) and Embeddings as a Service (EaaS). However, storing sensitive information as embeddings can be…

计算与语言 · 计算机科学 2024-06-06 Yiyi Chen , Heather Lent , Johannes Bjerva

Embeddings are functions that map raw input data to low-dimensional vector representations, while preserving important semantic information about the inputs. Pre-training embeddings on a large amount of unlabeled data and fine-tuning them…

机器学习 · 计算机科学 2020-08-21 Congzheng Song , Ananth Raghunathan

Conversational dense retrieval has shown to be effective in conversational search. However, a major limitation of conversational dense retrieval is their lack of interpretability, hindering intuitive understanding of model behaviors for…

信息检索 · 计算机科学 2024-06-04 Yiruo Cheng , Kelong Mao , Zhicheng Dou

The financial industry relies on deep learning models for making important decisions. This adoption brings new danger, as deep black-box models are known to be vulnerable to adversarial attacks. In computer vision, one can shape the output…

机器学习 · 计算机科学 2024-08-27 Alina Ermilova , Elizaveta Kovtun , Dmitry Berestnev , Alexey Zaytsev

Recently, video text detection, tracking, and recognition in natural scenes are becoming very popular in the computer vision community. However, most existing algorithms and benchmarks focus on common text cases (e.g., normal size, density)…

计算机视觉与模式识别 · 计算机科学 2023-12-06 Weijia Wu , Yiming Zhang , Yefei He , Luoming Zhang , Zhenyu Lou , Hong Zhou , Xiang Bai

The rapid advancement of speech generation models has heightened privacy and security concerns related to voice cloning (VC). Recent studies have investigated disrupting unauthorized voice cloning by introducing adversarial perturbations.…

声音 · 计算机科学 2025-07-04 Wei Fan , Kejiang Chen , Chang Liu , Weiming Zhang , Nenghai Yu

We design, implement, and evaluate adversarial decoding, a new, generic text generation technique that produces readable documents for different adversarial objectives. Prior methods either produce easily detectable gibberish, or cannot…

计算与语言 · 计算机科学 2025-03-07 Collin Zhang , Tingwei Zhang , Vitaly Shmatikov

Modern retrieval-augmented generation (RAG) systems convert sensitive content into high-dimensional embeddings and store them in vector databases that treat the resulting numerical artifacts as opaque. Major vector-store products do not…

密码学与安全 · 计算机科学 2026-05-14 Jascha Wanger

While text-to-image synthesis currently enjoys great popularity among researchers and the general public, the security of these models has been neglected so far. Many text-guided image generation models rely on pre-trained text encoders…

机器学习 · 计算机科学 2023-08-10 Lukas Struppek , Dominik Hintersdorf , Kristian Kersting

Deep learning models have achieved high performance on many tasks, and thus have been applied to many security-critical scenarios. For example, deep learning-based face recognition systems have been used to authenticate users to access many…

密码学与安全 · 计算机科学 2017-12-18 Xinyun Chen , Chang Liu , Bo Li , Kimberly Lu , Dawn Song

Dense retrieval systems have proven to be effective across various benchmarks, but require substantial memory to store large search indices. Recent advances in embedding compression show that index sizes can be greatly reduced with minimal…

信息检索 · 计算机科学 2026-01-16 L. Caspari , M. Dinzinger , K. Ghosh Dastidar , C. Fellicious , J. Mitrović , M. Granitzer

Word embedding has become ubiquitous and is widely used in various natural language processing (NLP) tasks, such as web retrieval, web semantic analysis, and machine translation, and so on. Unfortunately, training the word embedding in a…

计算与语言 · 计算机科学 2023-12-29 Wenting Li , Jiahong Xue , Xi Zhang , Huacan Chen , Zeyu Chen , Feijuan Huang , Yuanzhe Cai

Adversarial attacks expose vulnerabilities of deep learning models by introducing minor perturbations to the input, which lead to substantial alterations in the output. Our research focuses on the impact of such adversarial attacks on…

计算与语言 · 计算机科学 2023-09-14 Pavel Burnyshev , Elizaveta Kostenok , Alexey Zaytsev

We present Lit2Vec, a reproducible workflow for constructing and validating a chemistry corpus from the Semantic Scholar Open Research Corpus using conservative, metadata-based license screening. Using this workflow, we assembled an…

数据库 · 计算机科学 2026-04-15 Mahmoud Amiri , Jamile Mohammad Jafari , Sara Mostafapour , Thomas Bocklitz

Deep learning has gained tremendous success and great popularity in the past few years. However, deep learning systems are suffering several inherent weaknesses, which can threaten the security of learning models. Deep learning's wide use…

密码学与安全 · 计算机科学 2020-10-28 Yingzhe He , Guozhu Meng , Kai Chen , Xingbo Hu , Jinwen He

In the evolving landscape of text-to-image (T2I) diffusion models, the remarkable capability to generate high-quality images from textual descriptions faces challenges with the potential misuse of reproducing sensitive content. To address…

计算机视觉与模式识别 · 计算机科学 2024-07-24 Changhoon Kim , Kyle Min , Yezhou Yang