scholarly journals Generative models, language and active inference

Author(s):  
Karl Friston ◽  
Thomas Parr ◽  
Yan Yufik ◽  
Noor Sajid ◽  
Cathy J. Price ◽  
...  

This paper presents a biologically plausible generative model and inference scheme that is capable of simulating the generation and comprehension of language, when synthetic subjects 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 this 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 language processing 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.

Author(s):  
Norman D. Cook

Speech production in most people is strongly lateralized to the left hemisphere (LH), but language understanding is generally a bilateral activity. At every level of linguistic processing that has been investigated experimentally, the right hemisphere (RH) has been found to make characteristic contributions, from the processing of the affective aspects of intonation, through the appreciation of word connotations, the decoding of the meaning of metaphors and figures of speech, to the understanding of the overall coherency of verbal humour, paragraphs and short stories. If both hemispheres are indeed engaged in linguistic decoding and both processes are required to achieve a normal level of understanding, a central question concerns how the separate language functions on the left and right are integrated. This chapter reviews relevant studies on the hemispheric contributions to language processing and the role of interhemispheric communications in cognition.


1998 ◽  
Vol 172 (2) ◽  
pp. 142-146 ◽  
Author(s):  
Matthias Weisbrod ◽  
Sabine Maier ◽  
Sabine Harig ◽  
Ulrike Himmelsbach ◽  
Manfred Spitzer

BackgroundIn schizophrenia, disturbances in the development of physiological hemisphere asymmetry are assumed to play a pathogenetic role. The most striking difference between hemispheres is in language processing. The left hemisphere is superior in the use of syntactic or semantic information, whereas the right hemisphere uses contextual information more effectively.MethodUsing psycholinguistic experimental techniques, semantic associations were examined in 38 control subjects, 24 non-thought-disordered and 16 thought-disordered people with schizophrenia, for both hemispheres separately.ResultsDirect semantic priming did not differ between the hemispheres in any of the groups. Only thought-disordered people showed significant indirect semantic priming in the left hemisphere.ConclusionsThe results support: (a) a prominent role of the right hemisphere for remote associations; (b) enhanced spreading of semantic associations in thought-disordered subjects; and (c) disorganisation of the functional asymmetry of semantic processing in thought-disordered subjects.


2006 ◽  
Vol 15 (2) ◽  
pp. 123-140
Author(s):  
Robin Melrose

Robert Browning’s so-called dramatic lyric ‘My Last Duchess’ has been interpreted differently by different critics, some seeing the Duke as shrewd and others seeing him as witless. This article attempts to account for these differing interpretations by analysing indeterminacies in the language of the poem. Starting out with the work of Derrida on speech act theory, and findings on the role of the right hemisphere in language processing, it goes on to propose techniques of linguistic analysis based on systemic-functional linguistics and the concept of particle-waves of language first discussed in Melrose (1996). The article then analyses a number of these so-called particle-waves in ‘My Last Duchess’, and concludes that opposing interpretations of the Duke can be traced to the indeterminacies of language in the particle-waves.


2014 ◽  
Vol 1 (1) ◽  
pp. 57-76 ◽  
Author(s):  
Burcu Toğral Koca

Soğuk Savaş'ın sona ermesinin ardından, ulus ötesi etkileriyle küresel gelişmeler, güvenlik konusunda çalışan akademisyenleri, politika yapıcıları, siyasetçileri devlet-merkezli ve askeri-odaklı güvenlik tanımının ötesine taşımıştır. Bu dönüşümle, kalkınmayı güvenlikle ilişkilendiren söylemler de ivme kazanmıştır. Bu söylemlerin ve güvenlik ile kalkınma arasında kurulan bağın göç ve hareketlilik olguları üzerinde de önemli etkileri olmuştur. Zenginler, vasıflı işçiler, “gelişmiş” Batı’nın turistleri serbest dolaşım hakkından sorunsuz bir şekilde yararlanırken; aynı hakkı kullanmak isteyen, sığınmacıları, mültecileri, fakir, vasıfsız ve “düzensiz” göçmenleri kapsayan ve “artık nüfus” olarak nitelendirilebilecek gruplar yoğun ve sert teknolojik/bürokratik kontrol ve güvenlik pratiklerine maruz kalmaya başlamıştır. Bu pratiklerden bir tanesi, Avrupa Birliği’nin (AB) dış sınırlarını bu “istenmeyen”/”artık nüfusa” karşı korumak için faaliyete geçirilen Frontex’in kurulmasıdır. Operasyonel ve kurumsal yapısı itibariyle militarize bir organ olan Frontex, “artık nüfus”tan doğabilecek sözde tehditlere karşı bir kontrol teknolojisine dönüşmüştür. Bu çalışma, biyopolitikayı bir yönetim teknolojisi olarak ele alan Foucaultcu yaklaşım temelinde, bu değişim ve gelişmeleri eleştirel bir yolla çözümlemeyi amaçlamaktadır. Bunu yaparken, ilk olarak, AB’nin göç rejimini şekillendiren ve biyopolitikanın bir yansıması olan kalkınma ve güvenlik arasında kurulan bağ tahlil edilmiştir. Daha sonra, Frontex ve Frontex’in Türkiye-Yunanistan sınırındaki rolü incelenerek bu tahlil daha somut hale getirilmiştir. Son olarak, Frontex’in operasyonlarının göçmen haklarını ihlal ettiği ortaya konmuş ve bu nedenle AB ve ilgili devletler tarafından geliştirilen biyopolitik göç rejimine karşı insan odaklı, eleştirel bir siyasi duruşun geliştirilmesi gerektiği vurgulanmıştır. ENGLISH TITLE & ABSTRACTBiopolitics, security and the role of Frontex on the Turkish-Greek borderFollowing the end of the Cold War, global developments with their transnational effects have induced security scholars, policy makers and politicians to move beyond state-centric and military-focused conceptualization of security. In this transformation, discourses linking development to security have gained momentum. To put it differently, “liberal” states of the West have constructed a biopolitical distinction between “developed” and “underdeveloped” populations and administered the latter as a security threat to the former. Such framings and the nexus between security and development have had important repercussions for the mobility of people. The rich, skilled labours, tourists from “developed” West have come to enjoy the right to free movement without much of interruption. On the other hand, the “surplus” population, including asylum seekers, refugees, poor, unskilled and “irregular” migrants have been exposed to intense technological/bureaucratic control and surveillance practices. One of them is the introduction of Frontex for policing the European Union (EU)’s external borders against this “unwanted”/”surplus” population in conformity with new discourses linking development to security. This militarized body equipped with war-like devices has turned into a technology of containment related to the so-called threats stemming from “surplus” population. On the basis of these transformations, this paper aims to problematize and unpack these issues through building upon Foucauldian approaches on biopolitics as a technology of government. In particular, the paper, first, deconstructs the nexus between development and security in the EU’s migration regime. This analysis is made more tangible by looking into the activities of Frontex on the Turkish-Greek border. Finally, this paper draws the attention to the human rights implications of this security architecture and resultant practices. 


2013 ◽  
Vol 36 (3) ◽  
pp. 211-212 ◽  
Author(s):  
Thomas A. Farmer ◽  
Meredith Brown ◽  
Michael K. Tanenhaus

AbstractWe propose, following Clark, that generative models also play a central role in the perception and interpretation of linguistic signals. The data explanation approach provides a rationale for the role of prediction in language processing and unifies a number of phenomena, including multiple-cue integration, adaptation effects, and cortical responses to violations of linguistic expectations.


2021 ◽  
Vol 36 (3) ◽  
Author(s):  
Anjali Bhat ◽  
Thomas Parr ◽  
Maxwell Ramstead ◽  
Karl Friston

AbstractThere is a steadily growing literature on the role of the immune system in psychiatric disorders. So far, these advances have largely taken the form of correlations between specific aspects of inflammation (e.g. blood plasma levels of inflammatory markers, genetic mutations in immune pathways, viral or bacterial infection) with the development of neuropsychiatric conditions such as autism, bipolar disorder, schizophrenia and depression. A fundamental question remains open: why are psychiatric disorders and immune responses intertwined? To address this would require a step back from a historical mind–body dualism that has created such a dichotomy. We propose three contributions of active inference when addressing this question: translation, unification, and simulation. To illustrate these contributions, we consider the following questions. Is there an immunological analogue of sensory attenuation? Is there a common generative model that the brain and immune system jointly optimise? Can the immune response and psychiatric illness both be explained in terms of self-organising systems responding to threatening stimuli in their external environment, whether those stimuli happen to be pathogens, predators, or people? Does false inference at an immunological level alter the message passing at a psychological level (or vice versa) through a principled exchange between the two systems?


2021 ◽  
pp. 1-65
Author(s):  
Théophile Champion ◽  
Marek Grześ ◽  
Howard Bowman

Active inference is a state-of-the-art framework in neuroscience that offers a unified theory of brain function. It is also proposed as a framework for planning in AI. Unfortunately, the complex mathematics required to create new models can impede application of active inference in neuroscience and AI research. This letter addresses this problem by providing a complete mathematical treatment of the active inference framework in discrete time and state spaces and the derivation of the update equations for any new model. We leverage the theoretical connection between active inference and variational message passing as described by John Winn and Christopher M. Bishop in 2005. Since variational message passing is a well-defined methodology for deriving Bayesian belief update equations, this letter opens the door to advanced generative models for active inference. We show that using a fully factorized variational distribution simplifies the expected free energy, which furnishes priors over policies so that agents seek unambiguous states. Finally, we consider future extensions that support deep tree searches for sequential policy optimization based on structure learning and belief propagation.


2021 ◽  
Vol 15 ◽  
Author(s):  
Thomas Parr ◽  
Noor Sajid ◽  
Lancelot Da Costa ◽  
M. Berk Mirza ◽  
Karl J. Friston

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.


2021 ◽  
pp. 1-26
Author(s):  
Thijs van de Laar ◽  
Henk Wymeersch ◽  
İsmail Şenöz ◽  
Ayça Özçelikkale

Active inference (ActInf) is an emerging theory that explains perception and action in biological agents in terms of minimizing a free energy bound on Bayesian surprise. Goal-directed behavior is elicited by introducing prior beliefs on the underlying generative model. In contrast to prior beliefs, which constrain all realizations of a random variable, we propose an alternative approach through chance constraints, which allow for a (typically small) probability of constraint violation, and demonstrate how such constraints can be used as intrinsic drivers for goal-directed behavior in ActInf. We illustrate how chance-constrained ActInf weights all imposed (prior) constraints on the generative model, allowing, for example, for a trade-off between robust control and empirical chance constraint violation. Second, we interpret the proposed solution within a message passing framework. Interestingly, the message passing interpretation is not only relevant to the context of ActInf, but also provides a general-purpose approach that can account for chance constraints on graphical models. The chance constraint message updates can then be readily combined with other prederived message update rules without the need for custom derivations. The proposed chance-constrained message passing framework thus accelerates the search for workable models in general and can be used to complement message-passing formulations on generative neural models.


2021 ◽  
Author(s):  
Sebastian Rzydzik ◽  
Marcin Adamiec

This article describe a way to create generative model at the example of cargo bike model, which is very simple object which can be used to present all important rules applied during crating generative models. Great attention was paid to the issue of model parametrization which is elementary thing in all modelling. Besides these aspects, it is also shown how to transform parametric model into generative model using programming languages. In the last part of article was included tests of correct working of model which focused also to the right position cyclist on the bike and shows how model of cargo bike could change its sizes thanks to correctly created generative model.


Sign in / Sign up

Export Citation Format

Share Document