惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

S
SegmentFault 最新的问题
G
Google Developers Blog
Stack Overflow Blog
Stack Overflow Blog
WordPress大学
WordPress大学
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
罗磊的独立博客
月光博客
月光博客
IT之家
IT之家
爱范儿
爱范儿
Google DeepMind News
Google DeepMind News
小众软件
小众软件
C
Check Point Blog
B
Blog RSS Feed
H
Help Net Security
博客园 - 司徒正美
L
LangChain Blog
MongoDB | Blog
MongoDB | Blog
B
Blog
The Cloudflare Blog
Apple Machine Learning Research
Apple Machine Learning Research
Microsoft Security Blog
Microsoft Security Blog
M
MIT News - Artificial intelligence
N
Netflix TechBlog - Medium
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知

IBM Research

From error mitigation to fault-tolerant quantum computing | IBM Quantum Computing Blog A theoretical separation between quantum computers & LLMs Introducing IBM and NASA’s new foundation model for the Moon Switzerland's first IBM Quantum System Two | IBM Quantum Computing Blog Cleveland Clinic, RIKEN, IBM named Gordon Bell finalists | IBM Quantum Computing Blog How llm-d makes the most of the hardware you already have Ponder This Challenge - September 2026 - Loeschian Arithmetic Progressions IBM Quantum Nighthawk r2—more circuits, faster | IBM Quantum Computing Blog What happens when information theory accounts for reasoning? Granite 4.2 brings native reasoning to enterprise agents Qiskit Fermions: a modular toolbox for fermionic systems | IBM Quantum Computing Blog IBM’s new modular architecture for cryogenic systems | IBM Quantum Computing Blog QOBLIB: tracking progress in quantum optimization | IBM Quantum Computing Blog DocLang: a markup language for LLMs From vision to reality: a unified AI solver for the grid The search for quantum advantage in differential equations Ponder This Challenge - August 2026 - The Wheel of Buttons Quantum advantage through trusted quantum computation | IBM Quantum Computing Blog All of AI benchmarking at your fingertips What are spin qubits? | IBM Quantum Computing Blog IBM to acquire HRL Laboratories IBM commits $50M in quantum access for US Genesis Mission It’s time for cryptography to get its own abstraction layer It’s time for cryptography to get its own abstraction layer This could be the largest synthetic code dataset yet How to measure the performance of a quantum computer | IBM Quantum Computing Blog Release News: Qiskit v2.5 is here! | IBM Quantum Computing Blog CoFrGeNets replace the ‘bones’ of transformer-based models How training environments can teach AI models to misbehave What’s new at IBM Quantum - Q2 2026 | IBM Quantum Computing Blog
Ponder This Challenge - April 2026 - The Unlabeled Clock
Gadi Aleksandrowicz · 2026-04-01 · via IBM Research

The expected numerical solutions are:

  • 1541414
  • (30, 36, 40) => 3458432

This time, the phrasing of the puzzle, even after corrections, remained confusing to many solvers and we apologize for this. However, most wrong answers arose from the initial thinking that one should consider time to be discrete, looking only at moment of the form HH:MM. The riddle explicitly noted the clock's hands move continuously, and hence simple enumeration of moments of the form HH:MM is not sufficient to solve the riddle; a different approach is required.

We can model the current time as a value t[0,1)t\in[0,1) and the locations of the clock's hands become

  • H(t)=tH(t) = t
  • M(t)=12t (mod 1)M(t) = 12t \text{ }(\text{mod } 1)
  • S(t)=720t (mod 1)S(t) = 720t \text{ } (\text{mod } 1)

Let us define a three-component vector A(t)=(H(t),M(t),S(t))A(t)=(H(t), M(t), S(t)). Use the indices 1,2,31,2,3, i.e. A1(t)=H(t)A_1(t)=H(t), A2(t)=M(t)A_2(t)=M(t), A3(t)=S(t)A_3(t)=S(t).

Now, an ambigous moment arises from two distinct times t,ut, u such that we can permute the elements of A(t)A(t) and rotate them (i.e. add some rr modulo 1) and obtain A(s)A(s). Hence, let piS3pi\in S_3 be some permutation and r[0,1)r\in [0,1) be the rotation value, and we consider the system of three modular equations:

  • A1(u)=Aπ(1)(t)+r (mod 1)A_1(u) = A_{\pi(1)}(t)+r\text{ } (\text{mod } 1)
  • A2(u)=Aπ(2)(t)+r (mod 1)A_2(u) = A_{\pi(2)}(t)+r\text{ } (\text{mod } 1)
  • A3(u)=Aπ(3)(t)+r (mod 1)A_3(u) = A_{\pi(3)}(t)+r\text{ } (\text{mod } 1)

Subtracting the third equation from the first two, we obtain a system of two modular equations without rr:

  • A1(u)A3(u)=Aπ(1)(t)Aπ(3)(t) (mod 1)A_1(u)-A_3(u) = A_{\pi(1)}(t)-A_{\pi(3)}(t) \text{ } (\text{mod } 1)
  • A2(u)A3(u)=Aπ(2)(t)Aπ(3)(t) (mod 1)A_2(u)-A_3(u) = A_{\pi(2)}(t)-A_{\pi(3)}(t) \text{ } (\text{mod } 1)

Such modular equations can be solved using standard linear algebra. We demonstrate this in the case π(1)=2,π(2)=3,π(3)=1\pi(1)=2, \pi(2)=3, \pi(3)=1. In this case we obtain the modular equations

  • H(u)S(u)=M(t)H(t) (mod 1)H(u)-S(u) = M(t)-H(t) \text{ } (\text{mod } 1)
  • M(u)S(u)=S(t)H(t) (mod 1)M(u)-S(u) = S(t)-H(t) \text{ } (\text{mod } 1)

i.e.

  • u720u=12tt (mod 1)u-720u = 12t-t \text{ } (\text{mod } 1)
  • 12u720u=720tt (mod 1)12u-720u = 720t-t \text{ } (\text{mod } 1)

Which reduces to finding solutions to the linear system (11719719708)(tu)=(nm)\left(\begin{array}{cc} 11 & 719\\ 719 & 708 \end{array}\right)\left(\begin{array}{c} t\\ u \end{array}\right)=\left(\begin{array}{c} n\\ m \end{array}\right)

For n,mZn,m\in\mathbb{Z}.

This allows collection of all the possible ambiguous times. Situations like 03:00:00 and 09:00:00 still needs to be carefully ruled out.

The solution can be found manually; the goal of the "*" part was to encourage automation that will allow systematic checks for all (H,M,S) systems. However, some of those system yields infinite number of ambiguous moments; the intention was to find the maximum finite number.