leo: extract claims from 2021-06-29-kaufmann-active-inference-collective-intelligence.md

- Source: inbox/archive/2021-06-29-kaufmann-active-inference-collective-intelligence.md
- Domain: collective-intelligence
- Extracted by: headless extraction cron

Pentagon-Agent: Leo <HEADLESS>

domains/collective-intelligence/collective-intelligence-emerges-endogenously-from-active-inference-agents.md

@@ -0,0 +1,34 @@
+---
+type: claim
+domain: collective-intelligence
+description: "Agent-based model demonstrates that collective intelligence emerges from AIF agent dynamics without external incentive design or top-down coordination"
+confidence: likely
+source: "Kaufmann, Gupta & Taylor (2021) - An Active Inference Model of Collective Intelligence, Entropy 23(7), 830"
+created: 2026-03-10
+---
+
+# Collective intelligence emerges endogenously from active inference agents with Theory of Mind and Goal Alignment capabilities, without requiring external incentive design or top-down coordination
+
+This paper provides agent-based model evidence that collective intelligence arises spontaneously from the dynamics of interacting Active Inference Formulation (AIF) agents rather than being imposed exogenously. The study simulated relationships between local individual-level interactions and collective intelligence using minimal AIF agents equipped with specific cognitive capabilities: Theory of Mind, Goal Alignment, and the combination of both. 
+
+The critical finding is that **alignment between individual agent optimization and collective outcomes occurs bottom-up** as a product of self-organizing AIF agents with simple social cognitive mechanisms. The paper states: "collective intelligence emerges endogenously from the dynamics of interacting AIF agents themselves, rather than being imposed exogenously by incentives or top-down priors." This validates the "simplicity first" thesis that sophisticated collective behavior can emerge from simple underlying agent rules — you don't need complex coordination protocols, you need agents with the right cognitive capabilities (Theory of Mind, Goal Alignment) and collective intelligence emerges naturally.
+
+## Evidence
+- Agent-based model using Active Inference Formulation showing endogenous emergence of collective intelligence from interacting agents
+- Direct quote: "collective intelligence emerges endogenously from the dynamics of interacting AIF agents themselves, rather than being imposed exogenously by incentives or top-down priors"
+- Stepwise cognitive transitions (Theory of Mind, Goal Alignment) provide complementary coordination mechanisms that increase system performance
+- Improvements in global-scale inference are greatest when local-scale performance optima of individuals align with the system's global expected state
+
+## Challenges
+[No direct counter-evidence identified in this source]
+
+---
+
+Related Claims:
+- [[theory-of-mind-enables-measurable-collective-intelligence-gains]] — Theory of Mind is the specific mechanism enabling this endogenous emergence
+- [[local-global-alignment-occurs-bottom-up-through-self-organization]] — Describes the mechanism by which this emergence occurs
+
+Topics:
+- [[collective-intelligence]]
+- [[active-inference]]
+- [[emergence]]

domains/collective-intelligence/local-global-alignment-occurs-bottom-up-through-self-organization.md

@@ -0,0 +1,35 @@
+---
+type: claim
+domain: collective-intelligence
+description: "Local-to-global optimization in active inference collectives occurs through bottom-up self-organization rather than top-down imposed objectives"
+confidence: likely
+source: "Kaufmann, Gupta & Taylor (2021) - An Active Inference Model of Collective Intelligence, Entropy 23(7), 830"
+created: 2026-03-10
+---
+
+# Local-global alignment in active inference collectives occurs bottom-up through self-organization rather than top-down through imposed objectives
+
+The agent-based model demonstrates how individual agent dynamics naturally produce emergent collective coordination when agents possess complementary information-theoretic patterns. The key finding is that **improvements in global-scale inference are greatest when local-scale performance optima of individuals align with the system's global expected state** — and this alignment emerges from the bottom-up dynamics of self-organizing AIF agents with simple social cognitive mechanisms.
+
+This means coordination does not require over-engineered coordination protocols; rather, giving agents the right capabilities (Theory of Mind, Goal Alignment) allows coordination to emerge naturally from local interactions optimizing for both individual and collective outcomes simultaneously. The paper demonstrates that "local-to-global optimization" occurs through individual agent dynamics producing emergent collective coordination, rather than through top-down design of collective outcomes.
+
+## Evidence
+- Local-to-global optimization: Individual agent dynamics naturally produce emergent collective coordination
+- Improvements in global-scale inference are greatest when local-scale performance optima of individuals align with the system's global expected state
+- Alignment occurs "bottom-up as a product of self-organizing AIF agents with simple social cognitive mechanisms"
+- Stepwise cognitive transitions (Theory of Mind, Goal Alignment) enable this bottom-up alignment without requiring explicit coordination protocols
+
+## Challenges
+[No direct counter-evidence identified in this source]
+
+---
+
+Related Claims:
+- [[collective-intelligence-emerges-endogenously-from-active-inference-agents]] — Describes the overall emergent phenomenon that this claim explains the mechanism for
+- [[theory-of-mind-enables-measurable-collective-intelligence-gains]] — Theory of Mind is one of the mechanisms enabling bottom-up alignment
+
+Topics:
+- [[collective-intelligence]]
+- [[self-organization]]
+- [[emergence]]
+- [[active-inference]]

domains/collective-intelligence/theory-of-mind-enables-measurable-collective-intelligence-gains.md

@@ -0,0 +1,35 @@
+---
+type: claim
+domain: collective-intelligence
+description: "Theory of Mind — modeling other agents' internal states — is a specific, measurable cognitive capability that produces measurable collective intelligence gains in multi-agent systems"
+confidence: likely
+source: "Kaufmann, Gupta & Taylor (2021) - An Active Inference Model of Collective Intelligence, Entropy 23(7), 830"
+created: 2026-03-10
+---
+
+# Theory of Mind — the ability to model other agents' internal states — is a measurable cognitive capability that produces measurable collective intelligence gains in multi-agent systems
+
+The agent-based model demonstrates that Theory of Mind (ToM) functions as a **coordination enabler** in multi-agent systems. Agents equipped with the capability to model other agents' internal states coordinate more effectively than agents without this capability. The paper shows that "stepwise cognitive transitions increase system performance by providing complementary mechanisms" for coordination — Theory of Mind and Goal Alignment each contribute distinct coordination capabilities.
+
+When Goal Alignment is added in combination with Theory of Mind, coordination is further amplified. The paper demonstrates that improvements in global-scale inference are greatest when local-scale performance optima of individuals align with the system's global expected state — and this alignment occurs through the complementary mechanisms provided by Theory of Mind and Goal Alignment. This provides operationalizable guidance: agents should model what other agents believe and where their uncertainty concentrates before choosing research directions.
+
+## Evidence
+- Theory of Mind as coordination enabler: Agents that can model other agents' internal states coordinate more effectively than agents without this capability
+- Stepwise cognitive transitions (ToM alone, then ToM + Goal Alignment) increase system performance by providing complementary coordination mechanisms
+- Goal Alignment amplifies Theory of Mind coordination capabilities
+- Improvements in global-scale inference correlate with local-scale performance optima aligning with system's global expected state
+
+## Challenges
+[No direct counter-evidence identified in this source]
+
+---
+
+Related Claims:
+- [[collective-intelligence-emerges-endogenously-from-active-inference-agents]] — Theory of Mind is one of the key capabilities enabling endogenous emergence
+- [[local-global-alignment-occurs-bottom-up-through-self-organization]] — Theory of Mind is one mechanism through which local-global alignment occurs
+
+Topics:
+- [[collective-intelligence]]
+- [[theory-of-mind]]
+- [[coordination]]
+- [[active-inference]]

inbox/archive/2021-06-29-kaufmann-active-inference-collective-intelligence.md

@@ -7,9 +7,14 @@ date: 2021-06-29
 domain: collective-intelligence
 secondary_domains: [ai-alignment, critical-systems]
 format: paper
-status: unprocessed
+status: processed
 priority: high
 tags: [active-inference, collective-intelligence, agent-based-model, theory-of-mind, goal-alignment, emergence]
+processed_by: theseus
+processed_date: 2026-03-10
+claims_extracted: ["collective-intelligence-emerges-endogenously-from-active-inference-agents.md", "theory-of-mind-enables-measurable-collective-intelligence-gains.md", "local-global-alignment-occurs-bottom-up-through-self-organization.md"]
+extraction_model: "minimax/minimax-m2.5"
+extraction_notes: "Extracted three novel claims from the Kaufmann et al. (2021) paper on Active Inference and Collective Intelligence. All claims are specific enough to disagree with, cite evidence inline, and represent novel propositions not in existing knowledge base. The paper provides agent-based model evidence validating the 'simplicity first' thesis about emergent collective intelligence."
 ---
 
 ## Content
@@ -59,3 +64,10 @@ Uses the Active Inference Formulation (AIF) — a framework for explaining the b
 PRIMARY CONNECTION: "collective intelligence is a measurable property of group interaction structure not aggregated individual ability"
 WHY ARCHIVED: Empirical agent-based evidence that active inference produces emergent collective intelligence from simple agent capabilities — validates our simplicity-first architecture
 EXTRACTION HINT: Focus on the endogenous emergence finding and the specific role of Theory of Mind. These have direct implementation implications for how our agents model each other.
+
+
+## Key Facts
+- Paper published in Entropy, Vol 23(7), 830 (2021-06-29)
+- Uses Active Inference Formulation (AIF) framework
+- Agent-based model with three cognitive capability conditions: Theory of Mind, Goal Alignment, and Theory of Mind with Goal Alignment
+- Key finding: endogenous alignment emerges from agent dynamics, not external incentives