Fantastisk interessant studie som nevner at hjernen ikke benytter seg av input fra omgivelsene til å skape et bilde av omgivelsene, men til å bekrefte det bildet den allerede har skapt, altså bekrefte sin «fantasi» om virkeligheten.
Brains, it has recently been argued, are essentially prediction machines. They are bundles of cells that support perception and action by constantly attempting to match incoming sensory inputs with top-down expectations or predictions. This is achieved using a hierarchical generative model that aims to minimize prediction error within a bidirectional cascade of cortical processing. Such accounts offer a unifying model of perception and action, illuminate the functional role of attention, and may neatly capture the special contribution of cortical processing to adaptive success. This target article critically examines this “hierarchical prediction machine” approach, concluding that it offers the best clue yet to the shape of a unified science of mind and action. Sections 1 and 2 lay out the key elements and implications of the approach. Section 3 explores a variety of pitfalls and challenges, spanning the evidential, the methodological, and the more properly conceptual. The paper ends (sections 4 and 5) by asking how such approaches might impact our more general vision of mind, experience, and agency.
In this paradigm, the brain does not build its current model of distal causes (its model of how the world is) simply by accumulating, from the bottom-up, a mass of low-level cues such as edge-maps and so forth. Instead (see Hohwy 2007), the brain tries to predict the current suite of cues from its best models of the possible causes.
The Helmholtz Machine sought to learn new representations in a multilevel system (thus capturing increasingly deep regularities within a domain) without requiring the provision of copious pre-classified samples of the desired input/output mapping. In this respect, it aimed to improve (see Hinton 2010) upon standard back-propagation driven learning. It did this by using its own top-down connections to provide the desired states for the hidden units, thus (in effect) self-supervising the development of its perceptual “recognition model” using a generative model that tried to create the sensory patterns for itself (in “fantasy,” as it was sometimes said).3 (For a useful review of this crucial innovation and a survey of many subsequent developments, see Hinton 2007a).
Transposed (in ways we are about to explore) to the neural domain, this makes prediction error into a kind of proxy (Feldman & Friston 2010) for sensory information itself. Later, when we consider predictive processing in the larger setting of information theory and entropy, we will see that prediction error reports the “surprise” induced by a mismatch between the sensory signals encountered and those predicted.
A good place to start (following Rieke 1999) is with what might be thought of as the “view from inside the black box.” For, the task of the brain, when viewed from a certain distance, can seem impossible: it must discover information about the likely causes of impinging signals without any form of direct access to their source. Thus, consider a black box taking inputs from a complex external world. The box has input and output channels along which signals flow. But all that it “knows”, in any direct sense, are the ways its own states (e.g., spike trains) flow and alter. In that (restricted) sense, all the system has direct access to is its own states. The world itself is thus off-limits (though the box can, importantly, issue motor commands and await developments). The brain is one such black box.