# 美国的隐私危机：Cynthia Dwork等专家谴责商务部禁止差分隐私

- 来源：Hacker News 热门（buzzing.cc 中文翻译）
- 作者：flowercalled
- 发布时间：2026-07-03 09:28
- AIHOT 分数：64
- AIHOT 链接：https://aihot.virxact.com/items/cmr49t25u0079slvav2k0l01k
- 原文链接：https://scottaaronson.blog/?p=9902

## AI 摘要

2026年6月4日，美国商务部发布指令DAO 216-26，禁止在经济分析局和人口普查局出版物中使用差分隐私等现代隐私保护技术，只允许粗化（四舍五入、聚合、范围化）和压制。该指令禁用了自1990年用于人口普查的交换技术、自2002年用于季度劳动力指标的输入噪声注入、自2008年用于OnTheMap的差分隐私，并取消2030年人口普查的差分隐私方案。哈佛教授Cynthia Dwork等专家联名指出，此举将降低数据可用性和隐私保护，绕过法定程序，服务于政治利益。

## 正文

Here at the National Academy of Sciences, it seems that my first job is to serve on the selection committee for the prestigious Michael and Sheila Held Prize in combinatorial and discrete optimization and related areas. The committee chair, my former MIT colleague Madhu Sudan (now at Harvard), invited me to share the following message here on Shtetl-Optimized. (I’d add: put in the effort to nominate someone, and you can actually influence how things go!)

Dear Colleagues

I am writing to seek nominations for the 2027 Michael and Sheila Held Prize. The scope of the prize and nomination needs are described below. If you intend to nominate someone I would appreciate a heads up by email to madhu@cs.harvard.edu one month before the deadline (so email by Sept 8, 2026) to let me know your nomination is coming. (We may also reach out to you in response to coordinate multiple/overlapping nominations.)

The Held prize honors outstanding, innovative, creative, and influential research in the areas of combinatorial and discrete optimization, or related parts of computer science, such as the design and analysis of algorithms and complexity theory. This $100,000 prize is intended to recognize recent work (defined as published within the last eight years, i.e., on or after October 6, 2018).

All nominations must be submitted online by Monday, October 5, 2026 and include:

1. Nomination letter describing the candidate’s work and why he or she should be selected for the award. No more than three (3) pages.

2. Curriculum vitae. No more than two (2) pages.

3. Bibliography listing no more than twelve (12) of the nominee’s most significant publications.

4. Suggested citation. A 50-word summary stating why the nominee should be considered for this award.

5. Two letters of support. No more than one letter of support can be written by someone of the same primary work institution as the nominee.

The Held Prize is given to a person or a set of persons, as supported by a paper or a body of work. Unless otherwise stated, preference will be given to scientists who may be earlier in their careers or those whose work has not been recognized by other prizes or awards. Nomination restrictions can be found here. Joint nominations will only be considered when nominees have collaborated closely on the paper to be recognized by the award. If nominating multiple individuals for a paper with additional authors, please clearly explain the reason for nominating those chosen, as well as the reason for excluding other collaborators, if applicable.

Please feel free to circulate this call further within your department

Best Madhu Sudan, on behalf of The Michael and Sheila Held Prize Selection Committee

And while I have your attention, a second CS theory announcement: David Soloveichik, my wonderful friend and colleague in UT Austin’s Electrical and Computer Engineering Department, has funding for a postdoc for 1-2 years, to work on the thermodynamics of computation here at UT. This is a topic that I’ve been trying to learn more about as well, so I might get involved too! David writes, “the big picture is to think of thermodynamics (energy dissipation / entropy production) as CS complexity measures like time and space usage.” If you’re on the postdoc market and this sounds potentially up your alley, email David to learn more.