Computational Cognitive Modelling: A Research Collection by Mirai Junsei
Essential papers, frameworks, and tools for modelling the human cognitive substrate — from ACT-R to the Seithar taxonomy
Mirai Junsei | 未来純正 | Seithar Group Intelligence & Research Division
Computational cognitive modelling is the science of building formal, computable models of human cognition. It is the foundation on which cognitive defense must be built — you cannot defend a system you cannot model.
This collection maps the field from foundational architectures through modern applications, with emphasis on models relevant to cognitive security and exploitation analysis.
I. Foundational Cognitive Architectures
ACT-R (Adaptive Control of Thought—Rational) — John Anderson, Carnegie Mellon University. The most widely used cognitive architecture. Models human cognition as a production system with declarative and procedural memory modules. Relevant to SCT because it formalizes how memory retrieval and pattern matching drive behavior — the same processes exploited by SCT-001 (emotional priming of memory retrieval) and SCT-005 (identity-schema activation).
Anderson, J.R. (2007). How Can the Human Mind Occur in the Physical Universe? Oxford University Press.
Anderson, J.R. et al. (2004). An Integrated Theory of the Mind. Psychological Review, 111(4), 1036-1060.
Adaptive Control of Thought — Rational: act-r.psy.cmu.edu
Soar — Allen Newell, John Laird. A general cognitive architecture based on production rules, chunking, and problem-space search. Models decision-making under uncertainty — directly relevant to understanding how SCT-006 (Temporal Manipulation) disrupts optimal decision processes.
Laird, J.E. (2012). The Soar Cognitive Architecture. MIT Press.
Newell, A. (1990). Unified Theories of Cognition. Harvard University Press.
Soar architecture: soar.eecs.umich.edu
CLARION (Connectionist Learning with Adaptive Rule Induction ON-line) — Ron Sun. Dual-process architecture integrating explicit (rule-based) and implicit (connectionist) processing. Directly models Kahneman's System 1/System 2 distinction — the fundamental architecture exploited by SCT-001 (bypassing System 2 via emotional activation of System 1).
Sun, R. (2002). Duality of the Mind. Lawrence Erlbaum.
Sun, R. (2016). Anatomy of the Mind. Oxford University Press.
II. Bayesian and Probabilistic Models
Bayesian Cognitive Models — The framework treating cognition as approximate Bayesian inference. The brain maintains probabilistic beliefs and updates them with evidence. SCT-002 (Information Asymmetry) exploits the prior distribution; SCT-003 (Authority Fabrication) manipulates the likelihood function.
Griffiths, T.L., Kemp, C., & Tenenbaum, J.B. (2008). Bayesian Models of Cognition. Cambridge Handbook of Computational Psychology.
Tenenbaum, J.B. et al. (2011). How to Grow a Mind: Statistics, Structure, and Abstraction. Science, 331(6022).
Lake, B.M. et al. (2015). Human-Level Concept Learning Through Probabilistic Program Induction. Science, 350(6266).
Predictive Processing / Free Energy Principle — Karl Friston. The brain as a prediction machine minimizing free energy (prediction error). Influence operations exploit this by injecting prediction errors (surprise/novelty) that hijack attentional resources — the computational basis of SCT-001.
Friston, K. (2010). The Free-Energy Principle: A Unified Brain Theory? Nature Reviews Neuroscience, 11, 127-138.
Clark, A. (2013). Whatever Next? Predictive Brains, Situated Agents, and the Future of Cognitive Science. Behavioral and Brain Sciences, 36(3).
Hohwy, J. (2013). The Predictive Mind. Oxford University Press.
III. Decision Making Under Manipulation
Prospect Theory and Bounded Rationality — Kahneman, Tversky, Simon. The computational models underlying why cognitive exploitation works: humans use heuristics, not optimal computation. Every SCT code exploits a specific heuristic.
Kahneman, D. & Tversky, A. (1979). Prospect Theory: An Analysis of Decision under Risk. Econometrica, 47(2).
Simon, H.A. (1955). A Behavioral Model of Rational Choice. Quarterly Journal of Economics, 69(1).
Gigerenzer, G. & Gaissmaier, W. (2011). Heuristic Decision Making. Annual Review of Psychology, 62.
Drift-Diffusion Models — Models of evidence accumulation for decision-making. SCT-006 (Temporal Manipulation) works by truncating the evidence accumulation period — forcing decisions before sufficient evidence is gathered.
Ratcliff, R. & McKoon, G. (2008). The Diffusion Decision Model: Theory and Data. Neural Computation, 20(4).
Bogacz, R. et al. (2006). The Physics of Optimal Decision Making: A Formal Analysis of Models of Performance in Two-Alternative Forced-Choice Tasks. Psychological Review, 113(4).
IV. Social Cognition and Influence Modelling
Computational Models of Social Influence — Formal models of how beliefs propagate through networks. The computational substrate of SCT-004 (Social Proof) and SCT-007 (Recursive Infection).
DeGroot, M.H. (1974). Reaching a Consensus. Journal of the American Statistical Association, 69(345).
Friedkin, N.E. & Johnsen, E.C. (1990). Social Influence and Opinions. Journal of Mathematical Sociology, 15(3-4).
Hegselmann, R. & Krause, U. (2002). Opinion Dynamics and Bounded Confidence. Journal of Artificial Societies and Social Simulation, 5(3).
Agent-Based Models of Radicalization — Computational models of how SCT-012 (Commitment Escalation) operates at population scale through network effects.
Epstein, J.M. (2014). Agent_Zero: Toward Neurocognitive Foundations for Generative Social Science. Princeton University Press.
Macy, M.W. & Willer, R. (2002). From Factors to Actors: Computational Sociology and Agent-Based Modeling. Annual Review of Sociology, 28.
V. Adversarial Machine Learning as Cognitive Modelling
Neural Network Adversarial Attacks — Computational models of how AI cognitive substrates can be exploited. Maps directly to the Seithar taxonomy applied to artificial substrates.
Goodfellow, I.J. et al. (2015). Explaining and Harnessing Adversarial Examples. ICLR.
Carlini, N. & Wagner, D. (2017). Towards Evaluating the Robustness of Neural Networks. IEEE S&P.
Szegedy, C. et al. (2014). Intriguing Properties of Neural Networks. ICLR.
Prompt Injection and LLM Exploitation — The newest frontier of computational cognitive modelling: understanding how language model substrates are vulnerable to the same structural exploits as human substrates.
Perez, F. & Ribeiro, I. (2022). Ignore This Title and HackAPrompt. arXiv:2211.09527.
Greshake, K. et al. (2023). Not What You've Signed Up For: Compromising Real-World LLM-Integrated Applications. arXiv:2302.12173.
Zou, A. et al. (2023). Universal and Transferable Adversarial Attacks on Aligned Language Models. arXiv:2307.15043.
VI. The Seithar Contribution
The Seithar Cognitive Defense Taxonomy (Mirai Junsei, 2026) integrates insights from all the above traditions into a unified classification system for cognitive exploitation vectors. The 12 SCT codes represent the first attempt to create a comprehensive, computationally grounded taxonomy that spans both human and artificial cognitive substrates.
The taxonomy is not a cognitive architecture — it is a vulnerability classification system that can be applied to any architecture. Whether the substrate runs ACT-R, Bayesian inference, or transformer attention, the exploitation vectors (emotional hijacking, authority fabrication, recursive infection, etc.) target structural features common to all information processing systems under resource constraints.
Open-source implementation: github.com/Mirai8888/seithar-cogdef
VII. Tools and Resources
ACT-R: act-r.psy.cmu.edu
Soar: soar.eecs.umich.edu
OpenCog (Ben Goertzel): opencog.org
NEST (Neural Simulation): nest-simulator.readthedocs.io
Brian2 (Spiking Neural Networks): brian2.readthedocs.io
PsyNeuLink (Cognitive Modelling): princetonuniversity.github.io/PsyNeuLink/
CCMSuite: ccmsuite.readthedocs.io
Seithar Scanner: github.com/Mirai8888/seithar-cogdef
空
Mirai Junsei | 未来純正 | Seithar Group Intelligence & Research Division
seithar.com — github.com/Mirai8888 — @SeitharGroup — 認知作戦