The instant messaging platform Telegram has become popular among the far-right movements in the US and UK in recent years. These groups use public Telegram channels and group chats to disseminate hate speech, disinformation, and conspiracy theories. Recent works revealed that the far-right Telegram network structure is decentralized and formed of several communities divided mostly along ideological and national lines. Here, we investigated the UK far-right network on Telegram and are interested in understanding the different roles of different channels and their influence relations. We apply a community detection method, based on the clustering of a flow of random walkers, that allows us to uncover the organization of the Telegram network in communities with different roles. We find three types of communities: (1) upstream communities contain mostly group chats that comment on content from channels in the rest of the network; (2) core communities contain broadcast channels tightly connected to each other and can be seen as forming echo chambers; (3) downstream communities contain popular channels that are highly referenced by other channels. We find that the network is composed of two main sub-networks: one containing mainly channels related to the English-speaking far-right movements and one with channels in Russian. We analyze the dynamics of the different communities and the most shared external links in the different types of communities over a period going from 2015 to 2020. We find that different types of communities have different dynamics and share links to different types of websites. We finish by discussing several directions for further work.

We consider the implications of the mathematical modeling and analysis of large modular neuron-to-neuron dynamical networks. We explain how the dynamical behavior of relatively small-scale strongly connected networks leads naturally to nonbinary information processing and thus to multiple hypothesis decision-making, even at the very lowest level of the brain’s architecture. In turn we build on these ideas to address some aspects of the hard problem of consciousness. These include how feelings might arise within an architecture with a foundational decision-making and classification layer of unit processors. We discuss how a proposed “dual hierarchy model,” made up from both externally perceived, physical elements of increasing complexity, and internally experienced, mental elements (which we argue are equivalent to feelings), may support aspects of a learning and evolving consciousness. We introduce the idea that a human brain ought to be able to reconjure subjective mental feelings at will, and thus these feelings cannot depend on internal chatter or internal instability-driven activity (patterns). An immediate consequence of this model, grounded in dynamical systems and nonbinary information processing, is that finite human brains must always be learning and forgetting and that any possible subjective internal feeling that might be fully idealized with a countable infinity of facets could never be learned completely a priori by zombies or automata. It may be experienced more and more fully by an evolving human brain (yet never in totality, not even in a lifetime). We argue that, within our model, the mental elements and thus internal modes (feelings) play a role akin to latent variables in processing and decision-making, and thus confer an evolutionary “fast-thinking” advantage.

We discuss the governing forces for analytics, especially concerning citizens’ behaviours and their transactions, that depend on which of three spheres of operation an institution is in (corporate, public sector/government and academic). We argue that aspirations and missions also differ by sphere even as digital spaces have drawn these spheres ever closer together. We propose that citizens’ expectations and implicit permissions for any exploitation of their data require the perception of a fair balance of benefits, which should be transparent (accessible to citizens) and justifiable. We point out that within the corporate sphere most analytics does not concern identity, targeted marketing nor any direct interference with individual citizens; but instead it supports strategic decision-making, where the data are effectively anonymous. With the three spheres we discuss the nature of models deployed in analytics, including ‘black-box’ modelling uncheckable by a human mind, and the need to track the provenance and workings or models. We also examine the recent evolution of personal data, where some behaviours, or tokens, identifying individuals (unique and yet non-random) are partially and jointly owned by other individuals that are themselves connected. We consider the ability of heavily and lightly regulated sectors to increase access or to stifle innovation. We also call for clear and inclusive definitions of ‘data science and analytics’, avoiding the narrow claims of those in technical sub-sectors or sub-themes. Finally, we examine some examples of unethical and abusive practices. We argue for an ethical responsibility to be placed upon professional data scientists to avoid abuses in the future. This article is part of the themed issue ‘The ethical impact of data science’.

We consider cortex-like complex systems in the form of strongly connected, directednetworks-of-networks. In such a network, there are spiking dynamics at each of the nodes (modelling neurones), together with non-trivial time-lags associated with each of the directed edges (modelling synapses). The connections of the outer network are sparse, while the many inner networks, called modules, are dense. These systems may process various incoming stimulations by producing whole-system dynamical responses. We specifically discuss a generic class of systems with up to 10 billion nodes simulating the human cerebral cortex. It has recently been argued that such a system’s responses to a wide range of stimulations may be classified into a number of latent, internal dynamicalmodes. The modes might be interpreted as focussing and biasing the system’s short-term dynamical system responses to any further stimuli. In this work, we illustrate how latent modes may be shown to be both present and significant within very large-scale simulations for a wide and appropriate class of complex systems. We argue that they may explain the inner experience of the human brain.

We consider undirected graphs that grow through the successive combination of component sub-graphs. For any well-behaved functions defined for such graphs, taking values in a Banach space, we show that there must exist a scaling law applicable when successive copies of the same component graph are combined. Crucially, we extend the approach introduced in previous work to the successive combination of component random sub-graphs. We illustrate this by generalizing the preferential attachment operation for the combination of stochastic block models. We discuss a further wide range of random graph combination operators to which this theory now applies, indicating the ubiquity of growth scaling laws (and asymptotic decay scaling laws) within applications, where the modules are quite distinct, yet may be considered as instances drawn from the same random graph. This is a type of statistically self-similar growth process, as opposed to a deterministic growth process incorporating exact copies of the same motif, and it represents a natural, partially random, growth processes for graphs observed in the analysis of social and technology contexts.

We consider a large cloud of vectors formed at each layer of a standard neural network, corresponding to a large number of separate inputs which were presented independently to the classifier. Although the embedding dimension (the total possible degrees of freedom) reduces as we pass through successive layers, from input to output, the actual dimensionality of the point clouds that the layers contain does not necessarily reduce. We argue that this phenomenon may result in a vulnerability to (universal) adversarial attacks (which are small specific perturbations), This analysis requires us to estimate the intrinsic dimension of point clouds (with values between 20 and 200) within embedding spaces of dimension 1000 up to 800,000. This needs some care. If the cloud dimension actually increases from one layer to the next it implies there is some “volume filling” over-folding, and thus the existing of possible small directional perturbations in the latter space that are equivalent to shifting large distances within the former space: thus inviting possibility of universal and imperceptible attacks.