为什么数字签名能证明消息没被篡改？
Why can a digital signature prove a message wasn't altered?
发方用私钥对消息哈希签名，收方用公钥验证；任何篡改都会改变哈希使验证失败，结合单向哈希与公钥密码保证完整性。
The sender signs a hash of the message with a private key; the receiver verifies with the public key. Any tampering changes the hash and fails verification, combining one-way hashing and public-key crypto for integrity.
难度 专业 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Menezes、van Oorschot、Vanstone《Handbook of Applied Cryptography》（CRC Press）
Menezes, van Oorschot, Vanstone, Handbook of Applied Cryptography, CRC Press
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-1-4398-2191-6 (CRC Press)
#数字签名 #密码
为什么推荐系统本质上是矩阵补全？
Why is a recommender system essentially matrix completion?
用户乘物品的评分表大多空缺，推荐即预测空白格，用低秩矩阵分解把大表拆成小因子，是最优化与线性代数的应用。
A user-by-item rating table is mostly empty; recommending predicts the blanks by low-rank matrix factorization splitting the big table into small factors, an application of optimization and linear algebra.
难度 专业 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-11-04 · 2 条独立证据
来源 · Source
Resnick, Iacovou, Suchak & Bergstrom,《GroupLens》,Proceedings of the 1994 ACM conference on Computer supported,1994 (DOI: 10.1145/192844.192905)；Koren, Bell & Volinsky,《Matrix Factorization Techniques for Recommender Systems》,Computer,2009 (DOI: 10.1109/mc.2009.263)
Resnick, Iacovou, Suchak & Bergstrom, 'GroupLens', Proceedings of the 1994 ACM conference on Computer supported, 1994, DOI: 10.1145/192844.192905; Koren, Bell & Volinsky, 'Matrix Factorization Techniques for Recommender Systems', Computer, 2009, DOI: 10.1109/mc.2009.263
独立核验依据 · Independent verification: 核验依据 1 Crossref DOI 10.1109/mc.2009.263 · 核验依据 2 Crossref DOI 10.1145/192844.192905
#推荐系统 #矩阵
压缩感知（compressed sensing）为什么能用远少于“奈奎斯特”的采样就重建信号？
Why can compressed sensing reconstruct a signal from far fewer samples than the Nyquist rate demands?
压缩感知利用信号在某一基（如小波）下“稀疏”的特性：只要测量矩阵满足受限等距性质（RIP），用少量随机线性测量加 ℓ₁ 最小化，就能高概率恢复原信号。它颠覆了“采样率须≥2 倍最高频率”的直觉，使 MRI 等能在更少采样下成像。
Compressed sensing exploits a signal's sparsity in some basis (e.g., wavelets): if the measurement matrix satisfies the restricted isometry property (RIP), a small set of random linear measurements plus ℓ₁ minimization recovers the signal with high probability. It overturned the 'sample rate ≥ 2× bandwidth' intuition, enabling faster MRI and more.
难度 专业 · 适合 高中(15-18)
CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source
出处备注：著录含书名，但缺版次 / 出版社 / 年份中的一部分，需编辑补全后才能做独立比对。
Candès, E. J., 《压缩感知》, IEEE Transactions on Information Theory 52(2), 2006
Candès, E. J., 'Compressive Sampling', IEEE Trans. Inform. Theory 52(2), 2006
#压缩感知 #稀疏 #受限等距 #ℓ1最小化
为什么搜索引擎用PageRank排序网页？
Why do search engines rank pages with PageRank?
PageRank把网页当作图，用随机游走的概率分布衡量重要性：被重要页面链接多的页面更重要，本质是求解大规模矩阵的特征向量。
PageRank treats web pages as a graph and measures importance by a random-walk probability distribution: pages linked by important pages are more important. It is essentially finding the eigenvector of a huge matrix.
难度 专业 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-08-23 · 2 条独立证据
来源 · Source
吴军,《数学之美》第3版,人民邮电出版社,2020
Wu Jun, Beauty of Mathematics, 3rd ed., Posts & Telecom Press, 2020
独立核验依据 · Independent verification: 核验依据 2 ISBN 9787115537973（豆瓣书目条目: 数学之美）
#PageRank #矩阵
为什么卷积能识别图像里的边缘？
Why can convolution detect edges in images?
卷积用一个小滤波器（如差分核）扫过像素，突出相邻亮度的变化处即边缘；多层卷积堆叠学会从边缘到纹理再到物体，是CNN的核心运算。
Convolution sweeps a small filter (e.g. a difference kernel) over pixels, highlighting where brightness changes, i.e. edges. Stacking many convolutional layers learns from edges to textures to objects, the core of CNNs.
难度 专业 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Goodfellow, I., Bengio, Y. & Courville, A.，《Deep Learning》，MIT Press，2016
Goodfellow, I., Bengio, Y. & Courville, A., Deep Learning, MIT Press, 2016
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-0-262-03561-3 (MIT Press)
#卷积 #图像
零知识证明（zero-knowledge proof）为什么能让你“证明自己知道某个秘密”却不泄露秘密本身？
Why can a zero-knowledge proof let you 'prove you know a secret' without revealing anything about it?
零知识证明是一种交互协议：证明者通过多轮“挑战—应答”，让验证者确信某个命题为真（如“我知道某个离散对数”），却学不到秘密的任何信息（模拟器可造出同样分布的假记录）。它在区块链匿名认证与身份系统中至关重要，由戈德瓦塞尔、米卡利、拉克夫于 1985 年形式化。
A zero-knowledge proof is an interactive protocol where, through rounds of challenge–response, the verifier becomes convinced a claim is true (e.g., 'I know a discrete log') yet learns nothing about the secret (a simulator can fake transcripts with the same distribution). Vital to blockchain anonymous authentication, it was formalized by Goldwasser, Micali, and Rackoff in 1985.
难度 专业 · 适合 高中(15-18)
CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source
出处备注：著录含书名，但缺版次 / 出版社 / 年份中的一部分，需编辑补全后才能做独立比对。
Goldwasser, S., Micali, S., Rackoff, C., 《知识复杂性的零知识证明》, SIAM Journal on Computing 18(1), 1989
Goldwasser, S., Micali, S., Rackoff, C., 'The Knowledge Complexity of Interactive Proof Systems', SIAM J. Comput. 18(1), 1989
#零知识证明 #交互协议 #密码学 #隐私
为什么量子计算威胁现有密码？
Why does quantum computing threaten current cryptography?
秀尔算法能在多项式时间内分解大整数，使RSA失效；它利用量子叠加做并行计算。这推动了改用格等难解问题的后量子密码研究。
Shor's algorithm factors large integers in polynomial time, breaking RSA; it exploits quantum superposition for parallel computation. This drives post-quantum cryptography based on harder problems like lattices.
难度 专业 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-08-23 · 2 条独立证据
来源 · Source
吴军,《数学之美》第3版,人民邮电出版社,2020
Wu Jun, Beauty of Mathematics, 3rd ed., Posts & Telecom Press, 2020
独立核验依据 · Independent verification: 核验依据 2 ISBN 9787115537973（豆瓣书目条目: 数学之美）
#量子计算 #密码
为什么傅里叶变换能把声音拆成频率？
Why can the Fourier transform split sound into frequencies?
任何周期信号都能写成不同频率正弦波的叠加，傅里叶变换给出每个频率的强弱，是音频、图像、通信的数学基石。
Any periodic signal is a sum of sine waves at different frequencies; the Fourier transform gives each frequency's strength, the mathematical basis of audio, images and communications.
难度 专业 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-08-23 · 2 条独立证据
来源 · Source
吴军,《数学之美》第3版,人民邮电出版社,2020
Wu Jun, Beauty of Mathematics, 3rd ed., Posts & Telecom Press, 2020
独立核验依据 · Independent verification: 核验依据 2 ISBN 9787115537973（豆瓣书目条目: 数学之美）
#傅里叶 #信号
为什么哈希能快速查重又不泄露内容？
Why can a hash detect duplicates fast without revealing content?
哈希把任意数据映射成定长摘要，微小输入变化产生大不同输出（雪崩效应）。它单向、不可逆，用于校验与密码存储，是数论与位运算的应用。
A hash maps any data to a fixed-length digest where tiny input changes cause big output changes (avalanche effect). It is one-way and irreversible, used for checksums and password storage via number theory and bit operations.
难度 专业 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Menezes、van Oorschot、Vanstone《Handbook of Applied Cryptography》（CRC Press）
Menezes, van Oorschot, Vanstone, Handbook of Applied Cryptography, CRC Press
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-1-4398-2191-6 (CRC Press)
#哈希 #密码
为什么纠错码能让划伤的光盘还能读？
Why can error-correcting codes read a scratched disc?
里德-所罗门等纠错码在编码时加入冗余，划伤丢失的部分可由剩余信息数学重建，是有限域与编码论的成果。
Codes like Reed-Solomon add redundancy at encoding; scratched, missing parts are mathematically reconstructed from the rest, a result of finite fields and coding theory.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Menezes、van Oorschot、Vanstone《Handbook of Applied Cryptography》（CRC Press）
Menezes, van Oorschot, Vanstone, Handbook of Applied Cryptography, CRC Press
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-1-4398-2191-6 (CRC Press)
#纠错码 #编码
为什么压缩能变小又不丢信息（无损）？
Why can compression shrink data without losing information?
无损压缩利用统计冗余，例如把哈哈哈记为哈乘3，用更短编码表示高频模式。信息论（香农熵）给出了压缩的极限。
Lossless compression exploits statistical redundancy, e.g. writing hahaha as ha x3, using shorter codes for frequent patterns. Information theory (Shannon entropy) gives the compression limit.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-10-04 · 2 条独立证据
来源 · Source
《信息论基础》Cover 与 Thomas
Elements of Information Theory by Cover and Thomas
独立核验依据 · Independent verification: 核验依据 1 ISBN 9780471241959 (2006) · 核验依据 2 ISBN 9780665590955 (2002)
#压缩 #信息论
里德-所罗门码是什么？
What are Reed–Solomon codes?
里德-所罗门码是一类纠错码，用多项式在多点取值来编码，能纠正一段连续错误。它用于光盘、二维码和深空通信，使受损数据仍可恢复。
Reed–Solomon codes encode data as a polynomial sampled at points, correcting burst errors; used in CDs, QR codes and deep-space comms.
难度 进阶 · 适合 高中(15-18)
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Menezes、van Oorschot、Vanstone《Handbook of Applied Cryptography》（CRC Press）
Menezes, van Oorschot, Vanstone, Handbook of Applied Cryptography, CRC Press
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-1-4398-2191-6 (CRC Press)
#计算机数学 #纠错码
为什么GPS定位至少要3颗卫星，实际却用4颗？
Why does GPS need at least 3 satellites, but actually uses 4?
每颗卫星给出到它的距离（一个球面），3个球面交于一点定位；但还需要第4颗来校正接收机时钟误差（时间乘光速），所以实际用4颗。
Each satellite gives a distance (a sphere); three spheres intersect at a point. A fourth corrects the receiver's clock error (time times light speed), so four are actually used.
难度 进阶 · 适合 大众
CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source
出处备注：本题原挂出处与答案内容不符，已撤回；具体书目待补，暂不展示来源。
本题原挂出处与答案内容不符，已撤回；具体书目待补。
Previous citation withdrawn as unsupported for this answer; a specific source is pending.
#GPS #几何
为什么人工智能（神经网络）本质是数学？
Why is AI (neural networks) essentially mathematics?
神经网络是层层矩阵乘法加非线性函数，训练即用梯度下降最小化损失函数。所谓学习，其实是求解高维优化问题。
A neural network is layered matrix multiplication plus nonlinear functions; training minimizes a loss function by gradient descent. Learning is really solving a high-dimensional optimization problem.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
《深度学习》Goodfellow、Bengio 与 Courville
Deep Learning by Goodfellow Bengio and Courville
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-0-262-03561-3 (MIT Press)
#神经网络 #优化
为什么 JPEG 图像压缩要用离散余弦变换（DCT）把图像块转成频率？
Why does JPEG image compression use the discrete cosine transform (DCT) to turn image blocks into frequencies?
JPEG 把图像切成 8×8 小块，对每块做 DCT，把像素值变成一组“频率系数”：低频系数承载整体明暗、高频系数承载细微边缘。人眼对高频不敏感，于是可大幅量化并丢弃高频系数，再用之字形扫描与熵编码压缩。DCT 能量集中、实数运算快，且边界不引入明显振铃，因而取代傅里叶成为图像标准。
JPEG splits an image into 8×8 blocks and applies the DCT to each, converting pixels into frequency coefficients: low frequencies carry overall brightness, high frequencies carry fine edges. Since human eyes are insensitive to high frequencies, those coefficients are heavily quantized and dropped, then zig-zag scanned and entropy-coded. The DCT concentrates energy, uses fast real arithmetic, and avoids visible boundary ringing, so it became the image standard over the Fourier transform.
难度 进阶 · 适合 高中(15-18)
CC BY 4.0 · 已核验 · Verified 2026-11-04
来源 · Source
Wallace, G. K.,《The JPEG still picture compression standard》,IEEE Transactions on Consumer Electronics,1992 (DOI: 10.1109/30.125072)
Wallace, G. K., 'The JPEG still picture compression standard', IEEE Transactions on Consumer Electronics, 1992, DOI: 10.1109/30.125072
独立核验依据 · Independent verification: 核验依据 1 Crossref DOI 10.1109/30.125072
#离散余弦变换 #JPEG #图像压缩 #频率系数
快速傅里叶变换（FFT）是什么？
What is the Fast Fourier Transform (FFT)?
FFT 把离散傅里叶变换的计算量从 O(n²) 降到 O(n log n)，通过分治把信号分解为奇偶半。它是数字信号、图像压缩与卷积神经网络加速的底层引擎。
The FFT computes the discrete Fourier transform in O(n log n) via divide-and-conquer, powering digital signal processing, compression and CNNs.
难度 进阶 · 适合 高中(15-18)
CC BY 4.0 · 已核验 · Verified 2026-10-03 · 2 条独立证据
来源 · Source
《算法导论》Cormen 等
Introduction to Algorithms by Cormen Leiserson Rivest and Stein
独立核验依据 · Independent verification: 核验依据 1 ISBN 978-0-262-04630-5 · 核验依据 2 ISBN 9780691079516 (1957)
#计算机数学 #FFT
为什么区块链用哈希链防篡改？
Why does a blockchain use a hash chain to prevent tampering?
每个区块含上一块的哈希，改动任一区块会使后续所有哈希失效，需要重算全网，成本极高，从而保证不可篡改，是密码学与分布式的结合。
Each block contains the previous block's hash; altering any block invalidates all subsequent hashes, forcing a costly full recompute, thus ensuring immutability, a blend of cryptography and distributed systems.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-11-03
来源 · Source
中本聪《Bitcoin: A Peer-to-Peer Electronic Cash System》, 2008
S. Nakamoto, 'Bitcoin: A Peer-to-Peer Electronic Cash System', 2008
独立核验依据 · Independent verification: 核验依据 1 ARCHIVE 标识 academictorrents_8c271f4d2e92a3449e2d1bde633c
#区块链 #哈希
为什么密码学离不开质数？
Why does cryptography rely on prime numbers?
RSA等公钥加密基于大数易乘难分解：两个质数相乘很快，反过来分解极慢。质数的乘法单向性保证了通信安全。
Public-key encryption like RSA rests on easy-to-multiply-hard-to-factor: multiplying two primes is fast, factoring back is extremely slow. The one-way nature of prime multiplication secures communication.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-08-23 · 2 条独立证据
来源 · Source
吴军,《数学之美》第3版,人民邮电出版社,2020
Wu Jun, Beauty of Mathematics, 3rd ed., Posts & Telecom Press, 2020
独立核验依据 · Independent verification: 核验依据 2 ISBN 9787115537973（豆瓣书目条目: 数学之美）
#质数 #密码
为什么随机森林比单一模型更稳？
Why is a random forest more stable than a single model?
多个弱模型在不同样本上训练，投票平均可抵消各自误差（降低方差），是统计集成与自助采样思想的体现。
Many weak models trained on different samples vote or average, canceling individual errors (reducing variance). It reflects statistical ensemble and bootstrap ideas.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-10-04
来源 · Source
Efron, B. 与 Hastie, T., 《计算机时代的统计推断》, Cambridge University Press
Efron, B. & Hastie, T., 'Computer Age Statistical Inference', Cambridge University Press
独立核验依据 · Independent verification: 核验依据 1 ISBN 9788131503942 (2002)
#集成学习 #统计
为什么蒙特卡洛方法用随机就能算积分？
Why can Monte Carlo methods compute integrals using randomness?
用大量随机撒点，落在目标区域内的比例乘以总面积约等于积分值，样本越多越准。它把确定性问题转为概率估计，适合高维与复杂形状。
By scattering many random points, the fraction landing inside a region times the total area approximates the integral, more accurate with more samples. It turns a deterministic problem into a probabilistic estimate, great for high dimensions.
难度 进阶 · 适合 大众
CC BY 4.0 · 已核验 · Verified 2026-08-23 · 2 条独立证据
来源 · Source
吴军,《数学之美》第3版,人民邮电出版社,2020
Wu Jun, Beauty of Mathematics, 3rd ed., Posts & Telecom Press, 2020
独立核验依据 · Independent verification: 核验依据 2 ISBN 9787115537973（豆瓣书目条目: 数学之美）
#蒙特卡洛 #概率