Backdoor attacks and defenses targeting multi-domain ai models: A comprehensive review
Since the emergence of security concerns in artificial intelligence (AI), there has been
significant attention devoted to the examination of backdoor attacks. Attackers can utilize …
significant attention devoted to the examination of backdoor attacks. Attackers can utilize …
The unreasonable effectiveness of few-shot learning for machine translation
We demonstrate the potential of few-shot translation systems, trained with unpaired
language data, for both high and low-resource language pairs. We show that with only 5 …
language data, for both high and low-resource language pairs. We show that with only 5 …
Cater: Intellectual property protection on text generation apis via conditional watermarks
Previous works have validated that text generation APIs can be stolen through imitation
attacks, causing IP violations. In order to protect the IP of text generation APIs, recent work …
attacks, causing IP violations. In order to protect the IP of text generation APIs, recent work …
Spinning language models: Risks of propaganda-as-a-service and countermeasures
We investigate a new threat to neural sequence-to-sequence (seq2seq) models: training-
time attacks that cause models to “spin” their outputs so as to support an adversary-chosen …
time attacks that cause models to “spin” their outputs so as to support an adversary-chosen …
Safeguarding human values: rethinking US law for generative AI's societal impacts
Our interdisciplinary study examines the effectiveness of US law in addressing the complex
challenges posed by generative AI systems to fundamental human values, including …
challenges posed by generative AI systems to fundamental human values, including …