陶哲轩博客写数学问题时，通常先把对象定义清楚，再给直觉、反例和证明轮廓。把「Open question: deterministic UUP matrices」改写成可阅读的中文笔记，重点是：问题在问什么、已知到哪一步、下一步最容易走偏在哪。原站广告、分享条和导航已去掉。

问题在问什么

This problem in compressed sensing is an example of a derandomisation problem : take an object which, currently, can only be constructed efficiently by a probabilistic method, and figure out a deterministic construction of comparable strength and practicality. (For a general comparison of probabilistic and deterministic algorithms, I can point you to these slides by Avi Wigderson ).

I will define exactly what UUP matrices (the UUP stands for “ uniform uncertainty principle “) are later in this post. For now, let us just say that they are a generalisation of (rectangular) orthogonal matrices , in which the columns are locally almost orthogonal rather than globally perfectly orthogonal. Because of this, it turns out that one can pack significantly more columns into a UUP matrix than an orthogonal matrix, while still capturing many of the desirable features

已知结果和反例

There are several constructions of UUP matrices known today (e.g. random normalised Gaussian matrices, random normalised Bernoulli matrices, or random normalised minors of a discrete Fourier transform matrix) but (if one wants the sparsity parameter to be large) they are all probabilistic in nature; in particular, these constructions are not 100% guaranteed to actually produce a UUP matrix, although in many cases the failure rate can be proven to be exponentially small in the

Before we define UUP matrices explicitly, let us first recall what an (rectangular) orthogonal matrix is. We will view an matrix (m rows and n columns) as a collection of column vectors in the (complex) vector space , or equivalently as a means of linearly transformed an n-dimensional vector as an m-dimensional vector . I will call such a matrix orthogonal if these column vectors are orthonormal , i.e. they all have unit length and are orthogonal to each other: whenever .

证明或构造的主线

Orthonormal vectors have several pleasant properties. One of them is Pythagoras’ theorem

valid for all complex numbers . In other words, the linear encoding is an isometry. This implies that such an encoding can be inverted in a stable manner: given the encoded vector one can uniquely recover the original coefficients , and furthermore that small changes in w will not cause large fluctuations in . Indeed, one can reconstruct the coefficients quickly and explicitly by the formula

阅读时建议盯住的点

One would like to make n as large as possible, and m as small as possible, so that one can transform as high-dimensional vectors as possible using only as low-dimensional space as possible to store the transformed vectors. There is however a basic obstruction to this, which is that an orthogonal matrix can only exist when ; for if n is larger than m, then there are too many vectors to remain linearly independent in , and one must have a non-trivial linear independence

One can try to circumvent this restriction by weakening the condition (1) to (say)

阅读和落地时建议先做的 5 件事

1. 用自己的语言重写定义和结论，不看原文能不能说清对象是什么。
2. 找一个最小反例或边界情形，确认假设少一条会怎样。
3. 把证明拆成可独立检验的引理，每步只保留一个新想法。
4. 若涉及计算或形式化，先写可复现的小例子，再谈一般情形。
5. 记下尚未解决的缺口：缺估计、缺构造，还是缺正确的范畴。

和智能体、形式化工具怎么接

龙虾PRO做 OpenClaw 落地时，数学笔记最有用的部分往往是「可检验的步骤」：定义、反例、引理边界。智能体适合帮忙展开计算和检索，不适合代替你决定哪条假设能扔。

本文侧重全链路风控方法论。落地时请用自身业务单据做回放验证，不要把示例阈值直接当生产策略。 相关：风控体检 · 方案资源