A new computational private information retrieval (PIR) scheme based on random linear codes is presented. A matrix of messages from a McEliece scheme is used to query the server with carefully chosen errors. The server responds with the sum of the scalar multiple of the rows of the query matrix and the files. The user recovers the desired file by erasure decoding the response. Contrary to code-based cryptographic systems, the scheme presented here enables to use truly random codes, not only codes disguised as such. Further, we show the relation to the so-called error subspace search problem and quotient error search problem, which we assume to be difficult, and show that the scheme is secure against attacks based on solving these problems.

中文翻译：

---

基于计算代码的单服务器私有信息检索

提出了一种新的基于随机线性码的计算隐私信息检索（PIR）方案。来自 McEliece 方案的消息矩阵用于查询具有精心选择的错误的服务器。服务器以查询矩阵的行和文件的标量倍数的总和进行响应。用户通过对响应进行擦除解码来恢复所需的文件。与基于代码的密码系统相反，这里提出的方案能够使用真正的随机代码，而不仅仅是伪装成这样的代码。此外，我们展示了与所谓的错误子空间搜索问题和商错误搜索问题的关系，我们认为这些问题是困难的，并表明该方案在解决这些问题的基础上可以抵御攻击。