Search

The Free Energy Principle (FEP) states that under suitable conditions of weak coupling, random dynamical systems with sufficient degrees of freedom will behave so as to minimize an upper bound, formalized as a variational free energy, on surprisal (a.k.a., self-information). This upper bound can be read as a Bayesian prediction error. Equivalently, its negative is a lower bound on Bayesian model evidence (a.k.a., marginal likelihood). In short, certain random dynamical systems evince a kind of self-evidencing. Here, we reformulate the FEP in the formal setting of spacetime-background free, scale-free quantum information theory. We show how generic quantum systems can be regarded as observers, which with the standard freedom of choice assumption become agents capable of assigning semantics to observational outcomes. We show how such agents minimize Bayesian prediction error in environments characterized by uncertainty, insufficient learning, and quantum contextuality. We show that in its quantum-theoretic formulation, the FEP is asymptotically equivalent to the Principle of Unitarity. Based on these results, we suggest that biological systems employ quantum coherence as a computational resource and – implicitly – as a communication resource. We summarize a number of problems for future research, particularly involving the resources required for classical communication and for detecting and responding to quantum context switches. [ABSTRACT FROM AUTHOR]

Understanding information processing in the brain—and creating general-purpose artificial intelligence—are long-standing aspirations of scientists and engineers worldwide. The distinctive features of human intelligence are high-level cognition and control in various interactions with the world including the self, which are not defined in advance and are vary over time. The challenge of building human-like intelligent machines, as well as progress in brain science and behavioural analyses, robotics, and their associated theoretical formalisations, speaks to the importance of the world-model learning and inference. In this article, after briefly surveying the history and challenges of internal model learning and probabilistic learning, we introduce the free energy principle, which provides a useful framework within which to consider neuronal computation and probabilistic world models. Next, we showcase examples of human behaviour and cognition explained under that principle. We then describe symbol emergence in the context of probabilistic modelling, as a topic at the frontiers of cognitive robotics. Lastly, we review recent progress in creating human-like intelligence by using novel probabilistic programming languages. The striking consensus that emerges from these studies is that probabilistic descriptions of learning and inference are powerful and effective ways to create human-like artificial intelligent machines and to understand intelligence in the context of how humans interact with their world. [ABSTRACT FROM AUTHOR]

• New (hierarchical) generative model for linguistic interactions. • Builds on active inference formulations of dyadic interactions. • We simulate agents that ask and answer questions together. • Theta-gamma coupling emerges from belief updating under this framework. This paper presents a biologically plausible generative model and inference scheme that is capable of simulating communication between synthetic subjects who talk to each other. Building on active inference formulations of dyadic interactions, we simulate linguistic exchange to explore generative models that support dialogues. These models employ high-order interactions among abstract (discrete) states in deep (hierarchical) models. The sequential nature of language processing mandates generative models with a particular factorial structure—necessary to accommodate the rich combinatorics of language. We illustrate linguistic communication by simulating a synthetic subject who can play the 'Twenty Questions' game. In this game, synthetic subjects take the role of the questioner or answerer, using the same generative model. This simulation setup is used to illustrate some key architectural points and demonstrate that many behavioural and neurophysiological correlates of linguistic communication emerge under variational (marginal) message passing, given the right kind of generative model. For example, we show that theta-gamma coupling is an emergent property of belief updating, when listening to another. [ABSTRACT FROM AUTHOR]

This paper concerns the distributed intelligence or federated inference that emerges under belief-sharing among agents who share a common world—and world model. Imagine, for example, several animals keeping a lookout for predators. Their collective surveillance rests upon being able to communicate their beliefs—about what they see—among themselves. But, how is this possible? Here, we show how all the necessary components arise from minimising free energy. We use numerical studies to simulate the generation, acquisition and emergence of language in synthetic agents. Specifically, we consider inference, learning and selection as minimising the variational free energy of posterior (i.e., Bayesian) beliefs about the states, parameters and structure of generative models, respectively. The common theme—that attends these optimisation processes—is the selection of actions that minimise expected free energy, leading to active inference, learning and model selection (a.k.a., structure learning). We first illustrate the role of communication in resolving uncertainty about the latent states of a partially observed world, on which agents have complementary perspectives. We then consider the acquisition of the requisite language—entailed by a likelihood mapping from an agent's beliefs to their overt expression (e.g., speech)—showing that language can be transmitted across generations by active learning. Finally, we show that language is an emergent property of free energy minimisation, when agents operate within the same econiche. We conclude with a discussion of various perspectives on these phenomena; ranging from cultural niche construction, through federated learning, to the emergence of complexity in ensembles of self-organising systems. • Communication—and language in particular—is an emergent property of agents that seek evidence for generative models of their shared world. • Nested free energy minimising—evidence maximising—processes explain the emergence of language and its transmission over generations. • Reading these processes as inference integrates perspectives on communication; from generalised synchrony to cultural niche construction. [ABSTRACT FROM AUTHOR]

Bettinger, J. S., K. J. Friston. Conceptual Foundations of Physiological Regulation incorporating the Free Energy Principle & Self-Organized Criticality. NEUROSCI BIOBEHAV REV 23(x) 144-XXX, 2022. Since the late nineteen-nineties, the concept of homeostasis has been contextualized within a broader class of "allostatic" dynamics characterized by a wider-berth of causal factors including social, psychological and environmental entailments; the fundamental nature of integrated brain-body dynamics; plus the role of anticipatory, top-down constraints supplied by intrinsic regulatory models. Many of these evidentiary factors are integral in original descriptions of homeostasis; subsequently integrated; and/or cite more-general operating principles of self-organization. As a result, the concept of allostasis may be generalized to a larger category of variational systems in biology, engineering and physics in terms of advances in complex systems, statistical mechanics and dynamics involving heterogenous (hierarchical/heterarchical, modular) systems like brain-networks and the internal milieu. This paper offers a three-part treatment. 1) interpret "allostasis" to emphasize a variational and relational foundation of physiological stability; 2) adapt the role of allostasis as "stability through change" to include a "return to stability" and 3) reframe the model of homeostasis with a conceptual model of criticality that licenses the upgrade to variational dynamics. • Allostasis "allo" + "stasis" as "variational & relational stability" • Allostasis as "stability through change" return to stability (physiological resilience) • Regulatory physiological processing should also operate at/near critical phase transitions • Variational/relational stability of allostatic processing is associated with a Griffiths region • To minimize free energy is to self-organize internal states into a Griffiths regime [ABSTRACT FROM AUTHOR]

How do we navigate a deeply structured world? Why are you reading this sentence first – and did you actually look at the fifth word? This review offers some answers by appealing to active inference based on deep temporal models. It builds on previous formulations of active inference to simulate behavioural and electrophysiological responses under hierarchical generative models of state transitions. Inverting these models corresponds to sequential inference, such that the state at any hierarchical level entails a sequence of transitions in the level below. The deep temporal aspect of these models means that evidence is accumulated over nested time scales, enabling inferences about narratives (i.e., temporal scenes). We illustrate this behaviour with Bayesian belief updating – and neuronal process theories – to simulate the epistemic foraging seen in reading. These simulations reproduce perisaccadic delay period activity and local field potentials seen empirically. Finally, we exploit the deep structure of these models to simulate responses to local (e.g., font type) and global (e.g., semantic) violations; reproducing mismatch negativity and P300 responses respectively. [ABSTRACT FROM AUTHOR]

Predictive coding models, such as the ‘free-energy principle’ (FEP), have recently been discussed in relation to how interoceptive (afferent visceral feedback) signals update predictions about the state of the body, thereby driving autonomic mediation of homeostasis. This study appealed to ‘interoceptive inference’, under the FEP, to seek new insights into autonomic (dys)function and brain–body integration by examining the relationship between cardiac interoception and autonomic cardiac control in healthy controls and patients with forms of orthostatic intolerance (OI); to (i) seek empirical support for interoceptive inference and (ii) delineate if this relationship was sensitive to increased interoceptive prediction error in OI patients during head-up tilt (HUT)/symptom provocation. Measures of interoception and heart rate variability (HRV) were recorded whilst supine and during HUT in healthy controls (N = 20), postural tachycardia syndrome (PoTS, N = 20) and vasovagal syncope (VVS, N = 20) patients. Compared to controls, interoceptive accuracy was reduced in both OI groups. Healthy controls' interoceptive sensibility positively correlated with HRV whilst supine. Conversely, both OI groups' interoceptive awareness negatively correlated with HRV during HUT. Our pilot study offers initial support for interoceptive inference and suggests OI cohorts share a central pathophysiology underlying interoceptive deficits expressed across distinct cardiovascular autonomic pathophysiology. From a predictive coding perspective, OI patients' data indicates a failure to attenuate/modulate ascending interoceptive prediction errors, reinforced by the concomitant failure to engage autonomic reflexes during HUT. Our findings offer a potential framework for conceptualising how the human nervous system maintains homeostasis and how both central and autonomic processes are ultimately implicated in dysautonomia. [ABSTRACT FROM AUTHOR]

Given that eye movement control can be framed as an inferential process, how are the requisite forces generated to produce anticipated or desired fixation? Starting from a generative model based on simple Newtonian equations of motion, we derive a variational solution to this problem and illustrate the plausibility of its implementation in the oculomotor brainstem. We show, through simulation, that the Bayesian filtering equations that implement ‘planning as inference’ can generate both saccadic and smooth pursuit eye movements. Crucially, the associated message passing maps well onto the known connectivity and neuroanatomy of the brainstem – and the changes in these messages over time are strikingly similar to single unit recordings of neurons in the corresponding nuclei. Furthermore, we show that simulated lesions to axonal pathways reproduce eye movement patterns of neurological patients with damage to these tracts. [ABSTRACT FROM AUTHOR]

What we see is fundamentally dependent on where we look. Despite this seemingly obvious statement, many accounts of the neurobiology underpinning visual perception fail to consider the active nature of how we sample our sensory world. This review offers an overview of the neurobiology of visual perception, which begins with the control of saccadic eye movements. Starting from here, we can follow the anatomy backwards, to try to understand the functional architecture of neuronal networks that support the interrogation of a visual scene. Many of the principles encountered in this exercise are equally applicable to other perceptual modalities. For example, the somatosensory system, like the visual system, requires the sampling of data through mobile receptive epithelia. Analysis of a somatosensory scene depends on what is palpated, in much the same way that visual analysis relies on what is foveated. The discussion here is structured around the anatomical systems involved in active vision and visual scene construction, but will use these systems to introduce some general theoretical considerations. We will additionally highlight points of contact between the biology and the pathophysiology that has been proposed to cause a clinical disorder of scene construction – spatial hemineglect. [ABSTRACT FROM AUTHOR]

How do we navigate a deeply structured world? Why are you reading this sentence first – and did you actually look at the fifth word? This review offers some answers by appealing to active inference based on deep temporal models. It builds on previous formulations of active inference to simulate behavioural and electrophysiological responses under hierarchical generative models of state transitions. Inverting these models corresponds to sequential inference, such that the state at any hierarchical level entails a sequence of transitions in the level below. The deep temporal aspect of these models means that evidence is accumulated over nested time scales, enabling inferences about narratives (i.e., temporal scenes). We illustrate this behaviour with Bayesian belief updating – and neuronal process theories – to simulate the epistemic foraging seen in reading. These simulations reproduce perisaccadic delay period activity and local field potentials seen empirically. Finally, we exploit the deep structure of these models to simulate responses to local (e.g., font type) and global (e.g., semantic) violations; reproducing mismatch negativity and P300 responses respectively. [ABSTRACT FROM AUTHOR]

Tinnitus is a common disorder that often complicates hearing loss. Its mechanisms are incompletely understood. Current theories proposing pathophysiology from the ear to the cortex cannot individually – or collectively – explain the range of experimental evidence available. We propose a new framework, based on predictive coding, in which spontaneous activity in the subcortical auditory pathway constitutes a ‘tinnitus precursor’ which is normally ignored as imprecise evidence against the prevailing percept of ‘silence’. Extant models feature as contributory mechanisms acting to increase either the intensity of the precursor or its precision. If precision (i.e., postsynaptic gain) rises sufficiently then tinnitus is perceived. Perpetuation arises through focused attention, which further increases the precision of the precursor, and resetting of the default prediction to expect tinnitus. [ABSTRACT FROM AUTHOR]