All Things Predictive...
X-Spect's International (Virtual) Conference - November 4th-6th, 2020
In November we will be holding our final, and biggest, conference yet. And more, it will be completely in a virtual space. That means no matter where you are you will be able to attend!
In recent years the predictive processing paradigm has steadily gained influence in cognitive neuroscience and beyond. This framework casts the brain as a probabilistic prediction engine that continually generates predictions of the causal structure of the world in order to construct for itself (from the top down) incoming sensory signals. Conceiving of the brain in this way offers what many believe to be our first glimpse of a unified theory of mind and behaviour. That unified story applies not just to perception and action, but also to reasoning, planning, and emoting, and has implications for psychiatry, philosophy, anthropology and socio-cultural studies. It may also shed light on how to improve machine learning by adopting lessons from the biological brain. Critics worry, however, that the theory may be over-extending itself, securing scope at the expense of empirically distinctive content. In this, our final project conference, we will be celebrating, and sometimes questioning, All Things Predictive.
Schedule:
November 4th:
If you would like to attend please email us at [email protected] with "Conference" as the email title. We have a limited number of places for this event, so please contact as as soon as possible to secure your place.
See everyone in November!
In November we will be holding our final, and biggest, conference yet. And more, it will be completely in a virtual space. That means no matter where you are you will be able to attend!
In recent years the predictive processing paradigm has steadily gained influence in cognitive neuroscience and beyond. This framework casts the brain as a probabilistic prediction engine that continually generates predictions of the causal structure of the world in order to construct for itself (from the top down) incoming sensory signals. Conceiving of the brain in this way offers what many believe to be our first glimpse of a unified theory of mind and behaviour. That unified story applies not just to perception and action, but also to reasoning, planning, and emoting, and has implications for psychiatry, philosophy, anthropology and socio-cultural studies. It may also shed light on how to improve machine learning by adopting lessons from the biological brain. Critics worry, however, that the theory may be over-extending itself, securing scope at the expense of empirically distinctive content. In this, our final project conference, we will be celebrating, and sometimes questioning, All Things Predictive.
Schedule:
November 4th:
- 5-6pm GMT - Giovanni Pezzulo (CNR)
- Talk Title: Future-Oriented Cognition and the Predictive Brain.
- Predictive processing (PP) theories assume that the brain's generative model continuously generates predictions in the service of online action-perception and future-oriented forms of cognition. I will discuss how the predictive brain may support prospective functions, such as planning and imagination, by temporarily detaching from current action-perception loops. I will highlight two distinct but complementary usages of planning: ‘at decision time’, to support goal directed choices and sequential memory encoding, and ‘in the background’, to learn behavioural policies and to optimize internal models. I will exemplify this distinction by reviewing evidence on internally generated sequences of neuronal activity in the hippocampus during online navigation and offline periods; and link these neuronal processes to computational mechanisms of active inference.
- 6-7pm GMT - Lisa Feldman-Barrett (Northeastern)
- Talk Title: The Power of Prediction: An Emerging Paradigm for Studying Brain, Body and Mind
- The last two decades of neuroscience research has produced a growing number of studies suggesting that disparate psychological phenomena are all produced by domain-general predictive processes in the brain. When considered together, these studies form a coherent, neurobiologically-inspired research program for guiding psychological research about the mind and behavior. In this talk, we will consider several key assumptions and hypotheses that define a particular approach to predictive processing, with a focus on the role of metabolism, allostasis, and concepts.
- 5-6pm GMT - Paul Fletcher (Cambridge)
- Talk Title: Models, Predictions and Computational Psychiatry: Where Next?
- As part of its widely-endorsed candidature for the next grand theory of mind and brain, the predictive processing (PP) framework has demonstrated high degrees of face validity in developing plausible accounts of a wide array of psychological processes and psychiatric symptoms. But this is a relatively modest feat and there is growing criticism of an apparent contentment in the field with, among other things, offering re-descriptions of phenomena rather than truly providing a theoretical and empirically-validated framework for linking neurobiological processes to cognition and behaviour, function and dysfunction. I believe that PP may have a lot to offer and that some conceptual ground-clearing would be useful at this point. What do we mean by predictions? Are they necessarily “top-down”? How does Predictive Processing relate to Predictive Coding? Does any of this matter in applying the framework to the big questions? I think that these are crucial questions if PP is going to prove useful to me, as a psychiatrist, who sees it primarily as an over-arching framework within which to try to understand how mechanistic perturbations may create profound changes in the ways that a patient may come to perceive and comprehend the world.
- 6-7pm GMT - Abby Tabor (Bath)
- Talk Title: Disorder by Design: Embodied Active Inference in the City
- By 2050 68% of the world’s population will live in cities. With the ability to influence the daily interactions of two thirds of the world’s population, how we curate our cities and the built environment has the opportunity to be the first line treatment in preventative medicine. Yet, with 1 in 3 adults living with a chronic health condition, it seems our cities have come to neglect the human scale. In applying an embodied active inference approach to the way we shape our cities, can we establish a blueprint for health?
- Talk Title: Disorder by Design: Embodied Active Inference in the City
- 5-6pm GMT - Shaun Gallagher (Memphis)
- Talk Title: When You Know Something That Your Brain Doesn’t: Predictive Processing and Perceptual Illusions
- By showing how predictive processing explanations of perceptual illusions can be problematic, I argue that a strict internalist conception of PEM that takes the mind to be inferentially secluded from the non-neural body and environment is itself an illusion in which it only “seems to be more neurocentrically skull-bound than embodied or extended” (Hohwy 2013, emphasis added). Although an appeal to active inference will not solve this problem, it does point in the right direction, which is to consider the role of structural resistance tied to bodily and environmental processes.
- 6-7pm GMT - Nico Orlandi (UCSC)
- Talk Title: Representing Probabilities
- Influential research in perceptual and cognitive neuroscience offers probabilistic models of perceptual activity. The models trade in probabilistic representations and they describe perception as predicting unfolding sensory states in a way that roughly conforms to Bayesian inference. Should we adopt a realist attitude towards this type of research? And does this research introduce a puzzle for conscious experience? I reflect on what probabilistic representations are or would be, and on the seemingly strange status of perception. Nestled between sensory processing – which appears to be thoroughly probabilistic – and cognition – which also appears capable of expressing uncertainty, perceptual experience appears unitary and lacking in any uncertainty.
If you would like to attend please email us at [email protected] with "Conference" as the email title. We have a limited number of places for this event, so please contact as as soon as possible to secure your place.
See everyone in November!