Active inference as a theory of sentient behavior.

This review paper offers an overview of the history and future of active inference-a unifying perspective on action and perception. Active inference is based upon the idea that sentient behavior depends upon our brains' implicit use of internal models to predict, infer, and direct action. Our focus is upon the conceptual roots and development of this theory of (basic) sentience and does not follow a rigid chronological narrative. We trace the evolution from Helmholtzian ideas on unconscious inference, through to a contemporary understanding of action and perception. In doing so, we touch upon related perspectives, the neural underpinnings of active inference, and the opportunities for future development. Key steps in this development include the formulation of predictive coding models and related theories of neuronal message passing, the use of sequential models for planning and policy optimization, and the importance of hierarchical (temporally) deep internal (i.e., generative or world) models. Active inference has been used to account for aspects of anatomy and neurophysiology, to offer theories of psychopathology in terms of aberrant precision control, and to unify extant psychological theories. We anticipate further development in all these areas and note the exciting early work applying active inference beyond neuroscience. This suggests a future not just in biology, but in robotics, machine learning, and artificial intelligence.

Federated inference and belief sharing.

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.

Generating meaning: active inference and the scope and limits of passive AI.

Prominent accounts of sentient behavior depict brains as generative models of organismic interaction with the world, evincing intriguing similarities with current advances in generative artificial intelligence (AI). However, because they contend with the control of purposive, life-sustaining sensorimotor interactions, the generative models of living organisms are inextricably anchored to the body and world. Unlike the passive models learned by generative AI systems, they must capture and control the sensory consequences of action. This allows embodied agents to intervene upon their worlds in ways that constantly put their best models to the test, thus providing a solid bedrock that is - we argue - essential to the development of genuine understanding. We review the resulting implications and consider future directions for generative AI.

Bistable perception, precision and neuromodulation.

Bistable perception follows from observing a static, ambiguous, (visual) stimulus with two possible interpretations. Here, we present an active (Bayesian) inference account of bistable perception and posit that perceptual transitions between different interpretations (i.e. inferences) of the same stimulus ensue from specific eye movements that shift the focus to a different visual feature. Formally, these inferences are a consequence of precision control that determines how confident beliefs are and change the frequency with which one can perceive-and alternate between-two distinct percepts. We hypothesized that there are multiple, but distinct, ways in which precision modulation can interact to give rise to a similar frequency of bistable perception. We validated this using numerical simulations of the Necker cube paradigm and demonstrate the multiple routes that underwrite the frequency of perceptual alternation. Our results provide an (enactive) computational account of the intricate precision balance underwriting bistable perception. Importantly, these precision parameters can be considered the computational homologs of particular neurotransmitters-i.e. acetylcholine, noradrenaline, dopamine-that have been previously implicated in controlling bistable perception, providing a computational link between the neurochemistry and perception.

Path integrals, particular kinds, and strange things.

This paper describes a path integral formulation of the free energy principle. The ensuing account expresses the paths or trajectories that a particle takes as it evolves over time. The main results are a method or principle of least action that can be used to emulate the behaviour of particles in open exchange with their external milieu. Particles are defined by a particular partition, in which internal states are individuated from external states by active and sensory blanket states. The variational principle at hand allows one to interpret internal dynamics-of certain kinds of particles-as inferring external states that are hidden behind blanket states. We consider different kinds of particles, and to what extent they can be imbued with an elementary form of inference or sentience. Specifically, we consider the distinction between dissipative and conservative particles, inert and active particles and, finally, ordinary and strange particles. Strange particles can be described as inferring their own actions, endowing them with apparent autonomy or agency. In short-of the kinds of particles afforded by a particular partition-strange kinds may be apt for describing sentient behaviour.

A primer on Variational Laplace (VL).

This article details a scheme for approximate Bayesian inference, which has underpinned thousands of neuroimaging studies since its introduction 15 years ago. Variational Laplace (VL) provides a generic approach to fitting linear or non-linear models, which may be static or dynamic, returning a posterior probability density over the model parameters and an approximation of log model evidence, which enables Bayesian model comparison. VL applies variational Bayesian inference in conjunction with quadratic or Laplace approximations of the evidence lower bound (free energy). Importantly, update equations do not need to be derived for each model under consideration, providing a general method for fitting a broad class of models. This primer is intended for experimenters and modellers who may wish to fit models to data using variational Bayesian methods, without assuming previous experience of variational Bayes or machine learning. Accompanying code demonstrates how to fit different kinds of model using the reference implementation of the VL scheme in the open-source Statistical Parametric Mapping (SPM) software package. In addition, we provide a standalone software function that does not require SPM, in order to ease translation to other fields, together with detailed pseudocode. Finally, the supplementary materials provide worked derivations of the key equations.

Ecosystem bioelement variability is associated with freshwater animal aggregations at the aquatic-terrestrial interface.

The impacts of animals on the biogeochemical cycles of major bioelements like C, N, and P are well-studied across ecosystem types. However, more than 20 elements are necessary for life. The feedbacks between animals and the biogeochemical cycles of the other bioelements are an emerging research priority. We explored how much freshwater mussels (Bivalvia: Unionoida) were related to variability in ecosystem pools of 10 bioelements (Ca, Cu, Fe, K, Mn, Na, Mg, P, S and Zn) in streams containing a natural mussel density gradient in the US Interior Highlands. We studied the concentrations of these bioelements across the aquatic-terrestrial interface-in the porewater of riverine gravel bars, and the emergent macrophyte Justicia americana. Higher mussel density was associated with increased calcium in gravel bars and macrophytes. Mussel density also correlated with variability in iron and other redox-sensitive trace elements in gravel bars and macrophytes, although this relationship was mediated by sediment grain size. We found that two explanations for the patterns we observed are worthy of further research: (1) increased calcium availability in gravel bars near denser mussel aggregations may be a product of the buildup and dissolution of shells in the gravel bar, and (2) mussels may alter redox conditions, and thus elemental availability in gravel bars with fine sediments, either behaviorally or through physical structure provided by shell material. A better understanding of the physical and biogeochemical impacts of animals on a wide range of elemental cycles is thus necessary to conserve the societal value of freshwater ecosystems.

A Variational Synthesis of Evolutionary and Developmental Dynamics.

This paper introduces a variational formulation of natural selection, paying special attention to the nature of 'things' and the way that different 'kinds' of 'things' are individuated from-and influence-each other. We use the Bayesian mechanics of particular partitions to understand how slow phylogenetic processes constrain-and are constrained by-fast, phenotypic processes. The main result is a formulation of adaptive fitness as a path integral of phenotypic fitness. Paths of least action, at the phenotypic and phylogenetic scales, can then be read as inference and learning processes, respectively. In this view, a phenotype actively infers the state of its econiche under a generative model, whose parameters are learned via natural (Bayesian model) selection. The ensuing variational synthesis features some unexpected aspects. Perhaps the most notable is that it is not possible to describe or model a population of conspecifics per se. Rather, it is necessary to consider populations of distinct natural kinds that influence each other. This paper is limited to a description of the mathematical apparatus and accompanying ideas. Subsequent work will use these methods for simulations and numerical analyses-and identify points of contact with related mathematical formulations of evolution.

Understanding visual hallucinations: A new synthesis.

Despite decades of research, we do not definitively know how people sometimes see things that are not there. Eight models of complex visual hallucinations have been published since 2000, including Deafferentation, Reality Monitoring, Perception and Attention Deficit, Activation, Input, and Modulation, Hodological, Attentional Networks, Active Inference, and Thalamocortical Dysrhythmia Default Mode Network Decoupling. Each was derived from different understandings of brain organisation. To reduce this variability, representatives from each research group agreed an integrated Visual Hallucination Framework that is consistent with current theories of veridical and hallucinatory vision. The Framework delineates cognitive systems relevant to hallucinations. It allows a systematic, consistent, investigation of relationships between the phenomenology of visual hallucinations and changes in underpinning cognitive structures. The episodic nature of hallucinations highlights separate factors associated with the onset, persistence, and end of specific hallucinations suggesting a complex relationship between state and trait markers of hallucination risk. In addition to a harmonised interpretation of existing evidence, the Framework highlights new avenues of research, and potentially, new approaches to treating distressing hallucinations.

Cognitive effort and active inference.

This paper aims to integrate some key constructs in the cognitive neuroscience of cognitive control and executive function by formalising the notion of cognitive (or mental) effort in terms of active inference. To do so, we call upon a task used in neuropsychology to assess impulse inhibition-a Stroop task. In this task, participants must suppress the impulse to read a colour word and instead report the colour of the text of the word. The Stroop task is characteristically effortful, and we unpack a theory of mental effort in which, to perform this task accurately, participants must overcome prior beliefs about how they would normally act. However, our interest here is not in overt action, but in covert (mental) action. Mental actions change our beliefs but have no (direct) effect on the outside world-much like deploying covert attention. This account of effort as mental action lets us generate multimodal (choice, reaction time, and electrophysiological) data of the sort we might expect from a human participant engaging in this task. We analyse how parameters determining cognitive effort influence simulated responses and demonstrate that-when provided only with performance data-these parameters can be recovered, provided they are within a certain range.

Reward Maximization Through Discrete Active Inference.

Active inference is a probabilistic framework for modeling the behavior of biological and artificial agents, which derives from the principle of minimizing free energy. In recent years, this framework has been applied successfully to a variety of situations where the goal was to maximize reward, often offering comparable and sometimes superior performance to alternative approaches. In this article, we clarify the connection between reward maximization and active inference by demonstrating how and when active inference agents execute actions that are optimal for maximizing reward. Precisely, we show the conditions under which active inference produces the optimal solution to the Bellman equation, a formulation that underlies several approaches to model-based reinforcement learning and control. On partially observed Markov decision processes, the standard active inference scheme can produce Bellman optimal actions for planning horizons of 1 but not beyond. In contrast, a recently developed recursive active inference scheme (sophisticated inference) can produce Bellman optimal actions on any finite temporal horizon. We append the analysis with a discussion of the broader relationship between active inference and reinforcement learning.

In the Body's Eye: The computational anatomy of interoceptive inference.

A growing body of evidence highlights the intricate linkage of exteroceptive perception to the rhythmic activity of the visceral body. In parallel, interoceptive inference theories of affective perception and self-consciousness are on the rise in cognitive science. However, thus far no formal theory has emerged to integrate these twin domains; instead, most extant work is conceptual in nature. Here, we introduce a formal model of cardiac active inference, which explains how ascending cardiac signals entrain exteroceptive sensory perception and uncertainty. Through simulated psychophysics, we reproduce the defensive startle reflex and commonly reported effects linking the cardiac cycle to affective behaviour. We further show that simulated 'interoceptive lesions' blunt affective expectations, induce psychosomatic hallucinations, and exacerbate biases in perceptual uncertainty. Through synthetic heart-rate variability analyses, we illustrate how the balance of arousal-priors and visceral prediction errors produces idiosyncratic patterns of physiological reactivity. Our model thus offers a roadmap for computationally phenotyping disordered brain-body interaction.

What's special about space?

This commentary suggests that, although Markov blankets may have different interpretations in different systems, these distinctions rest not upon the type of blanket, but upon the model that determines the blanket. As an example, the conditions for a model in which the Markov blanket may be interpretable as a physical (spatial) boundary are considered.

Reclaiming saliency: Rhythmic precision-modulated action and perception.

Computational models of visual attention in artificial intelligence and robotics have been inspired by the concept of a saliency map. These models account for the mutual information between the (current) visual information and its estimated causes. However, they fail to consider the circular causality between perception and action. In other words, they do not consider where to sample next, given current beliefs. Here, we reclaim salience as an active inference process that relies on two basic principles: uncertainty minimization and rhythmic scheduling. For this, we make a distinction between attention and salience. Briefly, we associate attention with precision control, i.e., the confidence with which beliefs can be updated given sampled sensory data, and salience with uncertainty minimization that underwrites the selection of future sensory data. Using this, we propose a new account of attention based on rhythmic precision-modulation and discuss its potential in robotics, providing numerical experiments that showcase its advantages for state and noise estimation, system identification and action selection for informative path planning.

Bayesian Brains and the Rényi Divergence.

Under the Bayesian brain hypothesis, behavioral variations can be attributed to different priors over generative model parameters. This provides a formal explanation for why individuals exhibit inconsistent behavioral preferences when confronted with similar choices. For example, greedy preferences are a consequence of confident (or precise) beliefs over certain outcomes. Here, we offer an alternative account of behavioral variability using Rényi divergences and their associated variational bounds. Rényi bounds are analogous to the variational free energy (or evidence lower bound) and can be derived under the same assumptions. Importantly, these bounds provide a formal way to establish behavioral differences through an α parameter, given fixed priors. This rests on changes in α that alter the bound (on a continuous scale), inducing different posterior estimates and consequent variations in behavior. Thus, it looks as if individuals have different priors and have reached different conclusions. More specifically, α→0+ optimization constrains the variational posterior to be positive whenever the true posterior is positive. This leads to mass-covering variational estimates and increased variability in choice behavior. Furthermore, α→+∞ optimization constrains the variational posterior to be zero whenever the true posterior is zero. This leads to mass-seeking variational posteriors and greedy preferences. We exemplify this formulation through simulations of the multiarmed bandit task. We note that these α parameterizations may be especially relevant (i.e., shape preferences) when the true posterior is not in the same family of distributions as the assumed (simpler) approximate density, which may be the case in many real-world scenarios. The ensuing departure from vanilla variational inference provides a potentially useful explanation for differences in behavioral preferences of biological (or artificial) agents under the assumption that the brain performs variational Bayesian inference.

The evolution of brain architectures for predictive coding and active inference.

This article considers the evolution of brain architectures for predictive processing. We argue that brain mechanisms for predictive perception and action are not late evolutionary additions of advanced creatures like us. Rather, they emerged gradually from simpler predictive loops (e.g. autonomic and motor reflexes) that were a legacy from our earlier evolutionary ancestors-and were key to solving their fundamental problems of adaptive regulation. We characterize simpler-to-more-complex brains formally, in terms of generative models that include predictive loops of increasing hierarchical breadth and depth. These may start from a simple homeostatic motif and be elaborated during evolution in four main ways: these include the multimodal expansion of predictive control into an allostatic loop; its duplication to form multiple sensorimotor loops that expand an animal's behavioural repertoire; and the gradual endowment of generative models with hierarchical depth (to deal with aspects of the world that unfold at different spatial scales) and temporal depth (to select plans in a future-oriented manner). In turn, these elaborations underwrite the solution to biological regulation problems faced by increasingly sophisticated animals. Our proposal aligns neuroscientific theorising-about predictive processing-with evolutionary and comparative data on brain architectures in different animal species. This article is part of the theme issue 'Systems neuroscience through the lens of evolutionary theory'.