按年被提到
20202020 · 1 次2026
AI 安全、伦理与社会影响›模型窃取、记忆与水印·arXiv:2007.02220 · cs.CR

You Autocomplete Me: Poisoning Vulnerabilities in Neural Code Completion

Roei Schuster, Congzheng Song, Eran Tromer, Vitaly Shmatikov · 2020
开始阅读 ↗PDFalphaXiv 讨论
被提及 · 2 人首次 2020-07-13 · 最近 2021-10-27
2不同的人
2活跃月份
2讨论串
83被引(OpenAlex)
arXiv 摘要(候选层不写导读)

Code autocompletion is an integral feature of modern code editors and IDEs. The latest generation of autocompleters uses neural language models, trained on public open-source code repositories, to suggest likely (not just statically feasible) completions given the current context. We demonstrate that neural code autocompleters are vulnerable to poisoning attacks. By adding a few specially-crafted files to the autocompleter's training corpus (data poisoning), or else by directly fine-tuning the autocompleter on these files (model poisoning), the attacker can influence its suggestions for attacker-chosen contexts. For example, the attacker can "teach" the autocompleter to suggest the insecure ECB mode for AES encryption, SSLv3 for the SSL/TLS protocol version, or a low iteration count for password-based encryption. Moreover, we show that these attacks can be targeted: an autocompleter poisoned by a targeted attack is much more likely to suggest the insecure completion for files from a specific repo or specific developer. We quantify the efficacy of targeted and untargeted data- and model-poisoning attacks against state-of-the-art autocompleters based on Pythia and GPT-2. We then evaluate existing defenses against poisoning attacks and show that they are largely ineffective.

候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。

HN 书评

推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及
评论moyix · 2021-10

“This attack has been studied! https://arxiv.org/abs/2007.02220 Although our own work shows Copilot is pretty good at adding security flaws on its own: https://arxiv.org/abs/2108.09293”

GitHub Copilot available for JetBrains and Neovim原评论 ↗
被提交为 story · 1 次