Matches in SemOpenAlex for { <https://semopenalex.org/work/W4308643054> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W4308643054 abstract "Neural Machine Translation (NMT) systems have received much recent attention due to their human-level accuracy. While existing works mostly focus on either improving accuracy or testing accuracy robustness, the computation efficiency of NMT systems, which is of paramount importance due to often vast translation demands and real-time requirements, has surprisingly received little attention. In this paper, we make the first attempt to understand and test potential computation efficiency robustness in state-of-the-art NMT systems. By analyzing the working mechanism and implementation of 1455 public-accessible NMT systems, we observe a fundamental property in NMT systems that could be manipulated in an adversarial manner to reduce computation efficiency significantly. Our interesting observation is that the output length determines the computation efficiency of NMT systems instead of the input, where the output length depends on two factors: an often sufficiently large yet pessimistic pre-configured threshold controlling the max number of iterations and a runtime generated end of sentence (EOS) token. Our key motivation is to generate test inputs that could sufficiently delay the generation of EOS such that NMT systems would have to go through enough iterations to satisfy the pre-configured threshold. We present NMTSloth, which develops a gradient-guided technique that searches for a minimal and unnoticeable perturbation at character-level, token-level, and structure-level, which sufficiently delays the appearance of EOS and forces these inputs to reach the naturally-unreachable threshold. To demonstrate the effectiveness of NMTSloth, we conduct a systematic evaluation on three public-available NMT systems: Google T5, AllenAI WMT14, and Helsinki-NLP translators. Experimental results show that NMTSloth can increase NMT systems' response latency and energy consumption by 85% to 3153% and 86% to 3052%, respectively, by perturbing just one character or token in the input sentence. Our case study shows that inputs generated by NMTSloth significantly affect the battery power in real-world mobile devices (i.e., drain more than 30 times battery power than normal inputs)." @default.
- W4308643054 created "2022-11-13" @default.
- W4308643054 creator A5006956415 @default.
- W4308643054 creator A5009590736 @default.
- W4308643054 creator A5012605678 @default.
- W4308643054 creator A5016858094 @default.
- W4308643054 creator A5029676112 @default.
- W4308643054 date "2022-11-07" @default.
- W4308643054 modified "2023-09-25" @default.
- W4308643054 title "NMTSloth: understanding and testing efficiency degradation of neural machine translation systems" @default.
- W4308643054 cites W2562731582 @default.
- W4308643054 cites W2616028256 @default.
- W4308643054 cites W2799194071 @default.
- W4308643054 cites W2949128310 @default.
- W4308643054 cites W2954629067 @default.
- W4308643054 cites W2963125010 @default.
- W4308643054 cites W2963327228 @default.
- W4308643054 cites W2963393494 @default.
- W4308643054 cites W2963857521 @default.
- W4308643054 cites W2963993537 @default.
- W4308643054 cites W2970295111 @default.
- W4308643054 cites W2996851481 @default.
- W4308643054 cites W2998277219 @default.
- W4308643054 cites W3001434439 @default.
- W4308643054 cites W3034397670 @default.
- W4308643054 cites W3034866909 @default.
- W4308643054 cites W3035736465 @default.
- W4308643054 cites W3090988182 @default.
- W4308643054 cites W3091407209 @default.
- W4308643054 cites W3091441586 @default.
- W4308643054 cites W3101449015 @default.
- W4308643054 cites W3105599650 @default.
- W4308643054 cites W3163644064 @default.
- W4308643054 cites W4312648739 @default.
- W4308643054 doi "https://doi.org/10.1145/3540250.3549102" @default.
- W4308643054 hasPublicationYear "2022" @default.
- W4308643054 type Work @default.
- W4308643054 citedByCount "3" @default.
- W4308643054 countsByYear W43086430542022 @default.
- W4308643054 countsByYear W43086430542023 @default.
- W4308643054 crossrefType "proceedings-article" @default.
- W4308643054 hasAuthorship W4308643054A5006956415 @default.
- W4308643054 hasAuthorship W4308643054A5009590736 @default.
- W4308643054 hasAuthorship W4308643054A5012605678 @default.
- W4308643054 hasAuthorship W4308643054A5016858094 @default.
- W4308643054 hasAuthorship W4308643054A5029676112 @default.
- W4308643054 hasBestOaLocation W43086430541 @default.
- W4308643054 hasConcept C104317684 @default.
- W4308643054 hasConcept C113775141 @default.
- W4308643054 hasConcept C11413529 @default.
- W4308643054 hasConcept C119857082 @default.
- W4308643054 hasConcept C154945302 @default.
- W4308643054 hasConcept C185592680 @default.
- W4308643054 hasConcept C203005215 @default.
- W4308643054 hasConcept C38652104 @default.
- W4308643054 hasConcept C41008148 @default.
- W4308643054 hasConcept C45374587 @default.
- W4308643054 hasConcept C48145219 @default.
- W4308643054 hasConcept C55493867 @default.
- W4308643054 hasConcept C63479239 @default.
- W4308643054 hasConceptScore W4308643054C104317684 @default.
- W4308643054 hasConceptScore W4308643054C113775141 @default.
- W4308643054 hasConceptScore W4308643054C11413529 @default.
- W4308643054 hasConceptScore W4308643054C119857082 @default.
- W4308643054 hasConceptScore W4308643054C154945302 @default.
- W4308643054 hasConceptScore W4308643054C185592680 @default.
- W4308643054 hasConceptScore W4308643054C203005215 @default.
- W4308643054 hasConceptScore W4308643054C38652104 @default.
- W4308643054 hasConceptScore W4308643054C41008148 @default.
- W4308643054 hasConceptScore W4308643054C45374587 @default.
- W4308643054 hasConceptScore W4308643054C48145219 @default.
- W4308643054 hasConceptScore W4308643054C55493867 @default.
- W4308643054 hasConceptScore W4308643054C63479239 @default.
- W4308643054 hasLocation W43086430541 @default.
- W4308643054 hasLocation W43086430542 @default.
- W4308643054 hasOpenAccess W4308643054 @default.
- W4308643054 hasPrimaryLocation W43086430541 @default.
- W4308643054 hasRelatedWork W1967289688 @default.
- W4308643054 hasRelatedWork W1974878518 @default.
- W4308643054 hasRelatedWork W2375389409 @default.
- W4308643054 hasRelatedWork W2435130738 @default.
- W4308643054 hasRelatedWork W2935811960 @default.
- W4308643054 hasRelatedWork W2949968076 @default.
- W4308643054 hasRelatedWork W2961085424 @default.
- W4308643054 hasRelatedWork W2969407538 @default.
- W4308643054 hasRelatedWork W3099417250 @default.
- W4308643054 hasRelatedWork W4306674287 @default.
- W4308643054 isParatext "false" @default.
- W4308643054 isRetracted "false" @default.
- W4308643054 workType "article" @default.