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.

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.

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.

How Active Inference Could Help Revolutionise Robotics.

Recent advances in neuroscience have characterised brain function using mathematical formalisms and first principles that may be usefully applied elsewhere. In this paper, we explain how active inference-a well-known description of sentient behaviour from neuroscience-can be exploited in robotics. In short, active inference leverages the processes thought to underwrite human behaviour to build effective autonomous systems. These systems show state-of-the-art performance in several robotics settings; we highlight these and explain how this framework may be used to advance robotics.

Active Inference: Demystified and Compared.

Active inference is a first principle account of how autonomous agents operate in dynamic, nonstationary environments. This problem is also considered in reinforcement learning, but limited work exists on comparing the two approaches on the same discrete-state environments. In this letter, we provide (1) an accessible overview of the discrete-state formulation of active inference, highlighting natural behaviors in active inference that are generally engineered in reinforcement learning, and (2) an explicit discrete-state comparison between active inference and reinforcement learning on an OpenAI gym baseline. We begin by providing a condensed overview of the active inference literature, in particular viewing the various natural behaviors of active inference agents through the lens of reinforcement learning. We show that by operating in a pure belief-based setting, active inference agents can carry out epistemic exploration-and account for uncertainty about their environment-in a Bayes-optimal fashion. Furthermore, we show that the reliance on an explicit reward signal in reinforcement learning is removed in active inference, where reward can simply be treated as another observation we have a preference over; even in the total absence of rewards, agent behaviors are learned through preference learning. We make these properties explicit by showing two scenarios in which active inference agents can infer behaviors in reward-free environments compared to both Q-learning and Bayesian model-based reinforcement learning agents and by placing zero prior preferences over rewards and learning the prior preferences over the observations corresponding to reward. We conclude by noting that this formalism can be applied to more complex settings (e.g., robotic arm movement, Atari games) if appropriate generative models can be formulated. In short, we aim to demystify the behavior of active inference agents by presenting an accessible discrete state-space and time formulation and demonstrate these behaviors in a OpenAI gym environment, alongside reinforcement learning agents.

Degeneracy and Redundancy in Active Inference.

The notions of degeneracy and redundancy are important constructs in many areas, ranging from genomics through to network science. Degeneracy finds a powerful role in neuroscience, explaining key aspects of distributed processing and structure-function relationships in the brain. For example, degeneracy accounts for the superadditive effect of lesions on functional deficits in terms of a "many-to-one" structure-function mapping. In this paper, we offer a principled account of degeneracy and redundancy, when function is operationalized in terms of active inference, namely, a formulation of perception and action as belief updating under generative models of the world. In brief, "degeneracy" is quantified by the "entropy" of posterior beliefs about the causes of sensations, while "redundancy" is the "complexity" cost incurred by forming those beliefs. From this perspective, degeneracy and redundancy are complementary: Active inference tries to minimize redundancy while maintaining degeneracy. This formulation is substantiated using statistical and mathematical notions of degenerate mappings and statistical efficiency. We then illustrate changes in degeneracy and redundancy during the learning of a word repetition task. Finally, we characterize the effects of lesions-to intrinsic and extrinsic connections-using in silico disconnections. These numerical analyses highlight the fundamental difference between degeneracy and redundancy-and how they score distinct imperatives for perceptual inference and structure learning that are relevant to synthetic and biological intelligence.

Cancer Niches and Their Kikuchi Free Energy.

Biological forms depend on a progressive specialization of pluripotent stem cells. The differentiation of these cells in their spatial and functional environment defines the organism itself; however, cellular mutations may disrupt the mutual balance between a cell and its niche, where cell proliferation and specialization are released from their autopoietic homeostasis. This induces the construction of cancer niches and maintains their survival. In this paper, we characterise cancer niche construction as a direct consequence of interactions between clusters of cancer and healthy cells. Explicitly, we evaluate these higher-order interactions between niches of cancer and healthy cells using Kikuchi approximations to the free energy. Kikuchi's free energy is measured in terms of changes to the sum of energies of baseline clusters of cells (or nodes) minus the energies of overcounted cluster intersections (and interactions of interactions, etc.). We posit that these changes in energy node clusters correspond to a long-term reduction in the complexity of the system conducive to cancer niche survival. We validate this formulation through numerical simulations of apoptosis, local cancer growth, and metastasis, and highlight its implications for a computational understanding of the etiopathology of cancer.

Modules or Mean-Fields?

The segregation of neural processing into distinct streams has been interpreted by some as evidence in favour of a modular view of brain function. This implies a set of specialised 'modules', each of which performs a specific kind of computation in isolation of other brain systems, before sharing the result of this operation with other modules. In light of a modern understanding of stochastic non-equilibrium systems, like the brain, a simpler and more parsimonious explanation presents itself. Formulating the evolution of a non-equilibrium steady state system in terms of its density dynamics reveals that such systems appear on average to perform a gradient ascent on their steady state density. If this steady state implies a sufficiently sparse conditional independency structure, this endorses a mean-field dynamical formulation. This decomposes the density over all states in a system into the product of marginal probabilities for those states. This factorisation lends the system a modular appearance, in the sense that we can interpret the dynamics of each factor independently. However, the argument here is that it is factorisation, as opposed to modularisation, that gives rise to the functional anatomy of the brain or, indeed, any sentient system. In the following, we briefly overview mean-field theory and its applications to stochastic dynamical systems. We then unpack the consequences of this factorisation through simple numerical simulations and highlight the implications for neuronal message passing and the computational architecture of sentience.

Degeneracy in the neurological model of auditory speech repetition.

Both classic and contemporary models of auditory word repetition involve at least four left hemisphere regions: primary auditory cortex for processing sounds; pSTS (within Wernicke's area) for processing auditory images of speech; pOp (within Broca's area) for processing motor images of speech; and primary motor cortex for overt speech articulation. Previous functional-MRI (fMRI) studies confirm that auditory repetition activates these regions, in addition to many others. Crucially, however, contemporary models do not specify how regions interact and drive each other during auditory repetition. Here, we used dynamic causal modelling, to test the functional interplay among the four core brain regions during single auditory word and pseudoword repetition. Our analysis is grounded in the principle of degeneracy-i.e., many-to-one structure-function relationships-where multiple neural pathways can execute the same function. Contrary to expectation, we found that, for both word and pseudoword repetition, (i) the effective connectivity between pSTS and pOp was predominantly bidirectional and inhibitory; (ii) activity in the motor cortex could be driven by either pSTS or pOp; and (iii) the latter varied both within and between individuals. These results suggest that different neural pathways can support auditory speech repetition. This degeneracy may explain resilience to functional loss after brain damage.

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.

Generative Models for Active Vision.

The active visual system comprises the visual cortices, cerebral attention networks, and oculomotor system. While fascinating in its own right, it is also an important model for sensorimotor networks in general. A prominent approach to studying this system is active inference-which assumes the brain makes use of an internal (generative) model to predict proprioceptive and visual input. This approach treats action as ensuring sensations conform to predictions (i.e., by moving the eyes) and posits that visual percepts are the consequence of updating predictions to conform to sensations. Under active inference, the challenge is to identify the form of the generative model that makes these predictions-and thus directs behavior. In this paper, we provide an overview of the generative models that the brain must employ to engage in active vision. This means specifying the processes that explain retinal cell activity and proprioceptive information from oculomotor muscle fibers. In addition to the mechanics of the eyes and retina, these processes include our choices about where to move our eyes. These decisions rest upon beliefs about salient locations, or the potential for information gain and belief-updating. A key theme of this paper is the relationship between "looking" and "seeing" under the brain's implicit generative model of the visual world.

Simulating lesion-dependent functional recovery mechanisms.

Functional recovery after brain damage varies widely and depends on many factors, including lesion site and extent. When a neuronal system is damaged, recovery may occur by engaging residual (e.g., perilesional) components. When damage is extensive, recovery depends on the availability of other intact neural structures that can reproduce the same functional output (i.e., degeneracy). A system's response to damage may occur rapidly, require learning or both. Here, we simulate functional recovery from four different types of lesions, using a generative model of word repetition that comprised a default premorbid system and a less used alternative system. The synthetic lesions (i) completely disengaged the premorbid system, leaving the alternative system intact, (ii) partially damaged both premorbid and alternative systems, and (iii) limited the experience-dependent plasticity of both. The results, across 1000 trials, demonstrate that (i) a complete disconnection of the premorbid system naturally invoked the engagement of the other, (ii) incomplete damage to both systems had a much more devastating long-term effect on model performance and (iii) the effect of reducing learning capacity within each system. These findings contribute to formal frameworks for interpreting the effect of different types of lesions.

Active listening.

This paper introduces active listening, as a unified framework for synthesising and recognising speech. The notion of active listening inherits from active inference, which considers perception and action under one universal imperative: to maximise the evidence for our (generative) models of the world. First, we describe a generative model of spoken words that simulates (i) how discrete lexical, prosodic, and speaker attributes give rise to continuous acoustic signals; and conversely (ii) how continuous acoustic signals are recognised as words. The 'active' aspect involves (covertly) segmenting spoken sentences and borrows ideas from active vision. It casts speech segmentation as the selection of internal actions, corresponding to the placement of word boundaries. Practically, word boundaries are selected that maximise the evidence for an internal model of how individual words are generated. We establish face validity by simulating speech recognition and showing how the inferred content of a sentence depends on prior beliefs and background noise. Finally, we consider predictive validity by associating neuronal or physiological responses, such as the mismatch negativity and P300, with belief updating under active listening, which is greatest in the absence of accurate prior beliefs about what will be heard next.

Generative models, linguistic communication and active inference.

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.

Active inference on discrete state-spaces: A synthesis.

Active inference is a normative principle underwriting perception, action, planning, decision-making and learning in biological or artificial agents. From its inception, its associated process theory has grown to incorporate complex generative models, enabling simulation of a wide range of complex behaviours. Due to successive developments in active inference, it is often difficult to see how its underlying principle relates to process theories and practical implementation. In this paper, we try to bridge this gap by providing a complete mathematical synthesis of active inference on discrete state-space models. This technical summary provides an overview of the theory, derives neuronal dynamics from first principles and relates this dynamics to biological processes. Furthermore, this paper provides a fundamental building block needed to understand active inference for mixed generative models; allowing continuous sensations to inform discrete representations. This paper may be used as follows: to guide research towards outstanding challenges, a practical guide on how to implement active inference to simulate experimental behaviour, or a pointer towards various in-silico neurophysiological responses that may be used to make empirical predictions.

Paradoxical lesions, plasticity and active inference.

Paradoxical lesions are secondary brain lesions that ameliorate functional deficits caused by the initial insult. This effect has been explained in several ways; particularly by the reduction of functional inhibition, or by increases in the excitatory-to-inhibitory synaptic balance within perilesional tissue. In this article, we simulate how and when a modification of the excitatory-inhibitory balance triggers the reversal of a functional deficit caused by a primary lesion. For this, we introduce in-silico lesions to an active inference model of auditory word repetition. The first in-silico lesion simulated damage to the extrinsic (between regions) connectivity causing a functional deficit that did not fully resolve over 100 trials of a word repetition task. The second lesion was implemented in the intrinsic (within region) connectivity, compromising the model's ability to rebalance excitatory-inhibitory connections during learning. We found that when the second lesion was mild, there was an increase in experience-dependent plasticity that enhanced performance relative to a single lesion. This paradoxical lesion effect disappeared when the second lesion was more severe because plasticity-related changes were disproportionately amplified in the intrinsic connectivity, relative to lesioned extrinsic connections. Finally, this framework was used to predict the physiological correlates of paradoxical lesions. This formal approach provides new insights into the computational and neurophysiological mechanisms that allow some patients to recover after large or multiple lesions.

Neuromodulatory Control and Language Recovery in Bilingual Aphasia: An Active Inference Approach.

Understanding the aetiology of the diverse recovery patterns in bilingual aphasia is a theoretical challenge with implications for treatment. Loss of control over intact language networks provides a parsimonious starting point that can be tested using in-silico lesions. We simulated a complex recovery pattern (alternate antagonism and paradoxical translation) to test the hypothesis-from an established hierarchical control model-that loss of control was mediated by constraints on neuromodulatory resources. We used active (Bayesian) inference to simulate a selective loss of sensory precision; i.e., confidence in the causes of sensations. This in-silico lesion altered the precision of beliefs about task relevant states, including appropriate actions, and reproduced exactly the recovery pattern of interest. As sensory precision has been linked to acetylcholine release, these simulations endorse the conjecture that loss of neuromodulatory control can explain this atypical recovery pattern. We discuss the relevance of this finding for other recovery patterns.