首页|New Machine Learning Study Findings Recently Were Reported by Researchers at Geo rgia State University (A Semantic, Syntactic, and Context-aware Natural Language Adversarial Example Generator)
New Machine Learning Study Findings Recently Were Reported by Researchers at Geo rgia State University (A Semantic, Syntactic, and Context-aware Natural Language Adversarial Example Generator)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
2024 OCT 03 (NewsRx)-By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Research findings on Machine Learning are discussed in a new report. According to news reporting originating in Atlant a, Georgia, by NewsRx journalists, research stated, "Machine learning models are vulnerable to maliciously crafted Adversarial Examples (AEs). Training a machin e learning model with AEs improves its robustness and stability against adversar ial attacks." Funders for this research include National Science Foundation (NSF), Microsoft F aculty Fellowship Program. The news reporters obtained a quote from the research from Georgia State Univers ity, "It is essential to develop models that produce high-quality AEs. Developin g such models has been much slower in natural language processing (NLP) than in areas such as computer vision. This paper introduces a practical and efficient a dversarial attack model called SSCAE for Semantic, Syntactic, and Context-aware natural language AEs generator. SSCAE identifies important words and uses a mask ed language model to generate an early set of substitutions. Next, two well-know n language models are employed to evaluate the initial set in terms of semantic and syntactic characteristics. We introduce (1) a dynamic threshold to capture m ore efficient perturbations and (2) a local greedy search to generate high-quali ty AEs. As a black-box method, SSCAE generates humanly imperceptible and context -aware AEs that preserve semantic consistency and the source language's syntacti cal and grammatical requirements. The effectiveness and superiority of the propo sed SSCAE model are illustrated with fifteen comparative experiments and extensi ve sensitivity analysis for parameter optimization."
AtlantaGeorgiaUnited StatesNorth a nd Central AmericaCyborgsEmerging TechnologiesMachine LearningGeorgia St ate University
NETL
NSTL
