The Original Science Behind Eternal Recall
Eternal Recall isn't built on marketing claims. It's built on peer-reviewed cognitive science research spanning decades. Here are the foundational papers.
These papers describe the ACT-R cognitive architecture—the same memory models that Eternal Recall implements for AI. We provide both links to the original papers and accessible summaries.
Academic papers can be dense. We've written plain-English summaries of each paper's key contributions. Click the links to access the original research.
This paper introduces buffer decay—the mechanism that allows recently cleared memories to continue influencing context. It's a crucial extension to ACT-R that makes the architecture more psychologically plausible.
Buffer decay is what allows Eternal Recall to maintain broader context during conversations. Instead of only knowing what was just mentioned, the system maintains a fading awareness of recent context—exactly like human short-term memory.
Where L is time since clearing, d is decay rate (0.3-0.5)
Associations strengthen when chunks co-occur, weaken otherwise
This is an accessible tutorial that explains ACT-R to people outside the cognitive science community. It provides clear explanations of the core mechanisms that are scattered across the original ACT-R literature.
This paper provides the clearest explanation of the activation equations that Eternal Recall implements. The base-level learning equation (how memories strengthen and decay) and the spreading activation equation (how context affects recall) are directly implemented in our system.
"The activation A(t) of a knowledge chunk is the log-odds that, at time t, it will match to some production rule that ends up firing."
— From the paper, explaining what activation actually represents
Base-level activation plus spreading activation from context
Sum of decayed access events, plus constant
Sigmoid function determining recall success
These are the seminal works that established ACT-R as a cognitive architecture:
The definitive book on ACT-R architecture. Establishes the production system, declarative memory, and goal stack mechanisms.
Book - Lawrence Erlbaum Associates 📚 View on AmazonDescribes ACT-R as an integrated cognitive architecture with modular components. Published in Psychological Review.
Journal Article - Psychological Review 📄 PubMed 📄 ResearchGate (PDF)Introduces ACT-R and the "rational" optimization approach. Key text for understanding the theoretical foundations.
Book - Lawrence Erlbaum Associates 📚 View on AmazonExplains interference-based memory—why having more associations can slow retrieval. Published in Journal of Experimental Psychology.
Journal Article 📄 APA PsycNetApplies ACT-R to serial memory recall, modeling human error patterns in list learning.
Journal Article 📄 ACT-R ArchiveDerives the power law of learning from ACT-R principles. Key paper for understanding decay mathematics.
Journal Article 📄 ACT-R ArchiveOther papers that influenced Eternal Recall's design:
Godden, D., & Baddeley, A. (1975)
Seminal research on how environmental context affects recall. Famous study of divers learning underwater. Published in British Journal of Psychology.
📄 Semantic ScholarPeterson, L., & Peterson, M. J. (1959)
Established the 8-18 second duration of short-term memory without rehearsal. Foundation for buffer decay timing. Published in Journal of Experimental Psychology.
📄 PubMed 📖 SummaryWaugh, N. C., & Norman, D. A. (1965)
Distinction between primary (short-term) and secondary (long-term) memory systems. Published in Psychological Review.
📄 PubMed 📄 APA PsycNetCaporale, N., & Dan, Y. (2008)
Neuroscience research on Hebbian learning that inspired the associative learning mechanism. Published in Annual Review of Neuroscience.
📄 Scholarpedia Overview 📄 ScienceDirectMehlhorn, K., Taatgen, N.A., Lebiere, C., & Krems, J.F. (2011)
Application of ACT-R spreading activation to sequential diagnostic reasoning. Published in Journal of Experimental Psychology: Learning, Memory, & Cognition.
📄 PubMed 📥 Model Code & DataEternal Recall is built on these foundational research findings:
ACT-R Research: A(t) = B(t) + Σwjsij
In Eternal Recall: Memories have accessibility that changes based on use and context, following the original ACT-R mathematical framework.
ACT-R Research: B(t) = ln(Σtk-d)
In Eternal Recall: Memory accessibility follows a power law decay curve—fast initially, then stabilizing over time.
ACT-R Research: Σwjsij context term
In Eternal Recall: The Knowledge Graph implements spreading activation between related concepts and memories.
Research: Peterson & Peterson (1959) - short-term memory duration
In Eternal Recall: Recent discussion topics maintain contextual influence, based on cognitive science research.
Research: Hebbian plasticity principles
In Eternal Recall: The system learns relationships between concepts through repeated co-activation.
Research: Memory access strengthens retention
In Eternal Recall: Accessing memories reinforces them, keeping frequently-used information readily available.
Note: While Eternal Recall is built on published ACT-R research, our specific implementation—including custom adaptations and optimizations for AI memory systems—is proprietary.
Eternal Recall launches on Kickstarter in January 2026
Get Notified at Launch