Entropy
November 10, 2022
Gordana Dodig-Crnkovic
40 citations
Cognition is not unique to humans but is a property of all living organisms, from single cells upward. Viewed through an info-computational lens, structures in nature are information and their dynamics are computation from an agent's perspective. Cognition arises from networks of morphological and morphogenetic computations driven by self-assembly, self-organization, and autopoiesis. This article critiques the prevailing human-centric view of cognition, which faces unresolved problems, and reviews recent work on morphological computation, agency, basal cognition, and the free energy principle. It argues that older computational models, based on abstract symbol processing, ignored physical constraints and embodiment. Better understanding cognition is crucial for advancing artificial intelligence, robotics, and medicine.
Entropy
January 30, 2019
Aline Viol, Fernanda Palhano-Fontes, Heloisa Onias et al.
37 citations
A new network metric, geodesic entropy, measures the Shannon entropy of distances from one node to all others in a network, characterizing how much influence a node has based on the overall network structure. Applied to resting-state functional brain networks of humans, the metric differentiates ordinary consciousness from the altered state induced by Ayahuasca ingestion. On average, functional networks from subjects in the altered state show larger geodesic entropy than those in the ordinary state, suggesting the metric can reveal differences in brain network organization across states of consciousness.
Entropy
May 31, 2024
Michael Levin
31 citations
Memory is often studied for its ability to store and retrieve information faithfully, but this work argues that a more fundamental function is dynamically reinterpreting and modifying memories to fit an agent's changing self and environment. Drawing on examples from developmental biology, evolution, synthetic bioengineering, and neuroscience, the author proposes that memory preserves salience—what is relevant—rather than fidelity. This perspective applies across scales from cells to societies. The author suggests that continuous creative confabulation, from molecular to behavioral levels, resolves the persistence paradox for individuals and lineages. A processual view of life and mind implies that memories, as patterns in cognitive systems, can act as active agents in sense-making, supporting a view of life as nested perspectives engaged in polycomputation.
Entropy
January 22, 2024
Giulio Ruffini, Edmundo Lopez-Sola, Jakub Vohryzek et al.
15 citations
A framework called neural geometrodynamics, inspired by general relativity, describes how neural dynamics unfold at three timescales: fast (momentary activity), slow (synaptic plasticity), and ultraslow (metaplasticity). Psychedelics flatten the neural landscape, increasing entropy and complexity of fast dynamics, which disrupts functional integration. This destabilization counteracts pathological, rigid neural patterns by promoting fluid, adaptable states. The plasticity-enhancing effects of psychedelics amplify this shift, leading to acute systemic disorder and potentially longer-lasting increases in complexity that affect both short-term dynamics and long-term plastic processes, offering a holistic view of psychedelics' acute and lasting impacts.
Entropy
March 9, 2026
Jason Clarke
The Free Energy Principle and Active Inference explain how biological systems maintain organization under uncertainty but remain neutral on why there is experience at all. The Awareness-First Theory inverts the usual explanatory order by starting from the givenness of awareness itself and asking what must be the case for any world to appear coherently. This requirement is formalized as a Coherence Principle, expressed as a variational stationarity condition δA=0, which specifies the invariance of coherent awareness across changing appearances. Familiar variational principles like free-energy minimization (δF=0) and stationary-action physics (δS=0) can be understood as restricted projections of this parent constraint.