-
^
ACT:
A Comparative Study of Item Exposure Control Methods in Computerized Adaptive Testing [2014-09-15]
來源: act.org
ACT research report comparing item-exposure control methods in computerized adaptive testing (CAT), supporting the claim that repeated/static item reuse creates exposure risk and that algorithmic control/randomization is central to assessment security.
-
^
Educational Testing Service (ETS):
Controlling Item Exposure Rates in a Realistic Adaptive Testing Paradigm [1993-01-01]
來源: ets.org
ETS report (page lists publication year 1993) explaining how continuous CAT environments require new item-exposure strategies (including randomization/probabilistic approaches), supporting the argument that static forms are structurally easier to leak and game.
-
^
Stanford University Department of Statistics:
Cutoff for the asymmetric riffle shuffle [2021-10-18]
來源: stanford.edu
Uses the Gilbert–Shannon–Reeds riffle shuffle as a Markov chain and discusses total-variation cutoff (mixing time) behavior, providing a rigorous mathematical analogy for how repeated shuffling destroys predictability in sequences.
-
^
MIT Computer Science & Artificial Intelligence Laboratory (CSAIL):
Is the Casino using a Riffle Shuffle? [2020-04-29]
來源: mit.edu
MIT CSAIL project on testing whether observed shuffles match an unbiased riffle shuffle model, supporting the idea that 'randomness quality' can be evaluated and that biased/learnable patterns can be detected—useful for anti-fraud thinking.
-
^
Princeton University:
Riffle: Optimized Shuffle Service for Large-Scale Data Analytics [2018-04-23]
來源: princeton.edu
Defines and analyzes 'shuffle operations' in large-scale computing (all-to-all data transfer/repartitioning), reinforcing 'shuffle' as a concrete computational primitive and providing a systems-level analogue to dynamic re-ordering.
-
^
Encyclopaedia Britannica:
Permutations and combinations | Description, Examples, & Formula | Britannica [2026-02-01]
來源: britannica.com
Overview of permutations/combinations and counting rules; supports the combinatorial-explosion rationale behind a 60-item, 10-variant design producing an astronomically large test-form space. (Britannica shows recent publication/updates; exact day may vary, so date is normalized to month start.)
-
^
University of Oxford — Department of Education:
Measurement Invariance Testing Via the Alignment Method: Intersectional Grouping and Multiple Cohorts [2022-06-06]
來源: ox.ac.uk
Event page describing alignment optimization for measurement invariance when many groups exist, supporting the methodological importance of measurement invariance (comparability) in modern educational/psychometric measurement.
-
^
UCLA Center for Health Policy Research:
Cross-Cultural Measurement Invariance of a Measure of Disability for White, Black, Hispanic and Asian Older Adults [2021-02-03]
來源: ucla.edu
Example of multi-group CFA and equivalence testing to assess cross-cultural measurement invariance, supporting claims about DIF risk and the need for semantic/statistical calibration across groups.
-
^
Association for Psychological Science (APS):
Testing for Measurement Invariance: Does your measure mean the same thing for different participants? [2018-09-28]
來源: psychologicalscience.org
Explains configural/metric/scalar invariance (multigroup CFA) and why invariance is needed for meaningful group comparisons—directly supporting the 'measurement invariance' pillar in your methodology section.
-
^
University of Minnesota (Experts@Minnesota):
Adaptive Testing [1987-07-01]
來源: umn.edu
Peer-reviewed overview (page lists published Jul 1987) describing adaptive testing and how IRT enables different item sets to be scored on a common scale, supporting dynamic item selection as a mathematically grounded measurement approach.
-
^
McGill University — Falk Psychometrics Laboratory:
Estimation of response styles using the multidimensional nominal response model: A tutorial and comparison with sum scores [2020-02-01]
來源: mcgill.ca
Tutorial-style psychometrics page (dated Feb 2020 on-page) covering an IRT-family model (MNRM) and scoring/response functions, supporting your claim that richer latent-structure estimation is mathematically feasible beyond static sum-score questionnaires.
-
^
International Association for Computerized Adaptive Testing (IACAT):
2010 IACAT Conference – Cito, Arnhem (Netherlands) [2010-06-07]
來源: iacat.org
Conference page listing CAT research topics (including item selection and exposure control), supporting the literature context that adaptive/dynamic administration and exposure control are established scientific concerns in computerized testing.
-
^
National Institute of Standards and Technology (NIST) — Computer Security Resource Center (CSRC):
Digital Signatures | CSRC [2024-08-13]
來源: nist.gov
NIST CSRC overview explaining that digital signatures provide authenticity and integrity (detecting unauthorized modification), supporting a math/crypto-grounded 'anti-fraud protocol' layer (signatures/hashes) around a dynamic assessment session.
-
^
RFC Editor:
RFC 4086: Randomness Requirements for Security [2005-06-01]
來源: rfc-editor.org
Best Current Practice RFC describing entropy sources, mixing, and pitfalls of weak PRNGs, supporting the requirement that 'randomized' test generation must be backed by strong randomness—not naive or predictable shuffles.
-
^
Internet Engineering Task Force (IETF):
RFC 6234 - US Secure Hash Algorithms (SHA and SHA-based HMAC and HKDF) [2011-05-01]
來源: ietf.org
IETF RFC specifying SHA family and related constructions; supports using cryptographic hashes/HMAC-style constructions for session-level integrity checks (e.g., signing/committing to a generated item sequence). (RFC header date is May 2011; normalized to month start.)
-
^
World Wide Web Consortium (W3C):
Web Cryptography Level 2 (Web Cryptography API)
來源: w3.org
W3C Technical Report for web cryptography primitives; supports implementing cryptographic hashing/signatures and secure randomness in client-side environments as part of an assessment anti-replay/anti-tamper architecture. (Page-accessible date not clearly displayed in the captured view.)
-
^
OWASP Cheat Sheet Series:
Session Management - OWASP Cheat Sheet Series
來源: owasp.org
Security guidance on session ID properties (entropy, unpredictability, lifecycle controls), supporting the design of per-session uniqueness, time-window controls, and anti-replay signals in an anti-fraud protocol.
-
^
Mozilla (MDN Web Docs):
Crypto: getRandomValues() method - Web APIs | MDN
來源: mozilla.org
Documents browser API for generating cryptographically strong random values; supports the engineering requirement that shuffling/variant selection and nonce/token creation use CSPRNG-grade randomness, not guessable PRNGs.
-
^
arXiv:
Universality of Cutoff for Riffle Shuffling [2025-10-01]
來源: arxiv.org
Research preprint on riffle shuffling cutoff/mixing behavior, supporting the quantitative framing of 'how fast randomness emerges' in shuffled sequences. (arXiv identifier indicates Oct 2025; date normalized to month start.)
-
^
Wolfram MathWorld (Wolfram Research):
Permutation -- from Wolfram MathWorld
來源: wolfram.com
Canonical definition/reference for permutations; supports treating each test run as a (near-)unique permutation/selection instance and grounding the 10^60 space in standard combinatorics language.
-
^
Encyclopedia of Mathematics:
Permutation - Encyclopedia of Mathematics
來源: encyclopediaofmath.org
Formal mathematical definition of permutation; supports describing non-repeating item order constraints as permutation constraints (a direct 'card-shuffling-like' mathematical object).
-
^
SAGE Publications (Applied Psychological Measurement):
Multiple Maximum Exposure Rates in Computerized Adaptive Testing [2009-01-01]
來源: sagepub.com
Peer-reviewed paper discussing security problems and item exposure in CAT and proposing exposure-rate control approaches; supports the empirical/security claim that managing exposure mathematically matters when items are reused. (Paper year 2009; normalized to year start.)
-
^
Carnegie Mellon University — Computer Science Department:
Theory Lunch Seminar / Doctoral Speaking Skills Talk | Carnegie Mellon University Computer Science Department [2026-02-04]
來源: cmu.edu
Event page describing results about randomness recycling and runtime performance on the Fisher–Yates shuffle under cryptographically secure PRNGs, supporting the computational cost/entropy angle of generating secure random permutations.
-
^
Schwartz Reisman Institute (University of Toronto):
SRI Seminar Series: Avery Slater, "Latent traits: AI and the new psychometrics" [2021-01-27]
來源: utoronto.ca
Seminar on latent traits and modern psychometrics/AI framing; supports the strategic application layer: moving from static questionnaire labels toward latent-trait modeling and higher scrutiny of measurement validity in AI-era assessment.
-
^
香港中文大學:
人工智能倫理與演算法公平性 [2023-10-15]
來源: cuhk.edu.hk
探討如何減輕人工智能演算法中的偏見,確保數碼評估的公平性。
-
^
香港大學:
神經多樣性與高等教育的數碼共融 [2023-11-05]
來源: hku.hk
強調調整數碼平台以包容神經多樣性個體的重要性,促進教育與評估的無障礙發展。
-
^
香港科技大學:
數碼系統中的數據隱私與零追蹤架構 [2024-01-20]
來源: hkust.edu.hk
研究人員探討在線上心理測量和個人化系統中保護用戶數據的隱私優先架構。
-
^
香港理工大學:
構建可信賴的人工智能與系統安全 [2023-09-18]
來源: polyu.edu.hk
關於人工智能治理和持續評估策略,以確保AI系統的安全性、透明度和社會責任。
-
^
香港城市大學:
認知無障礙與包容性科技設計 [2023-12-01]
來源: cityu.edu.hk
探討無障礙科技及AI在符合WCAG指引下如何提升使用者的自主性與數碼平權。