Seeing social interactions

Trends in Cognitive Sciences, Journal Year: 2023, Volume and Issue: 27(12), P. 1165 - 1179

Published: Oct. 5, 2023

Seeing the interactions between other people is a critical part of our everyday visual experience, but recognizing social others often considered outside scope vision and grouped with higher-level cognition like theory mind. Recent work, however, has revealed that recognition efficient automatic, well modeled by bottom-up computational algorithms, occurs in visually-selective regions brain. We review recent evidence from these three methodologies (behavioral, computational, neural) converge to suggest core interaction perception visual. propose framework for how this process carried out brain offer directions future interdisciplinary investigations perception.

Language: Английский

---

EEG Signal Processing Techniques and Applications—2nd Edition

Sensors, Journal Year: 2025, Volume and Issue: 25(3), P. 805 - 805

Published: Jan. 29, 2025

Electroencephalography (EEG), as a well-established, non-invasive tool, has been successfully applied to wide range of conditions due its many evident advantages, such economy, portability, easy operation, accessibility, and widespread availability in hospitals [...]

Language: Английский

---

Multilevel development of cognitive abilities in an artificial neural network

Proceedings of the National Academy of Sciences, Journal Year: 2022, Volume and Issue: 119(39)

Published: Sept. 19, 2022

Several neuronal mechanisms have been proposed to account for the formation of cognitive abilities through postnatal interactions with physical and sociocultural environment. Here, we introduce a three-level computational model information processing acquisition abilities. We propose minimal architectural requirements build these levels, how parameters affect their performance relationships. The first sensorimotor level handles local nonconscious processing, here during visual classification task. second or globally integrates from multiple processors via long-ranged connections synthesizes it in global, but still nonconscious, manner. third cognitively highest consciously. It is based on global workspace (GNW) theory referred as conscious level. use trace delay conditioning tasks to, respectively, challenge levels. Results highlight necessity epigenesis selection stabilization synapses at both scales allow network solve two tasks. At scale, dopamine appears necessary properly provide credit assignment despite temporal between perception reward. level, presence interneurons becomes maintain self-sustained representation within GNW absence sensory input. Finally, while balanced spontaneous intrinsic activity facilitates scales, excitatory/inhibitory ratio increases performance. discuss plausibility neurodevelopmental artificial intelligence terms.

Language: Английский

---

The Darkside of Artificial Intelligence and the Metaverse in Scientific Research and Publishing

Advances in computational intelligence and robotics book series, Journal Year: 2025, Volume and Issue: unknown, P. 275 - 304

Published: April 24, 2025

This chapter explores the ethical, legal, and societal risks of Artifificial Intelligence (AI) Metaverse in scientific research publishing. While AI aids data analysis peer review, it perpetuating biases that could distort findings compromise integrity. The Metaverse, as a new digital space for academic engagement, introduces challenges like privacy, intellectual property concerns, opportunities fraud. Furthermore, algorithmic publishing amplify visibility disparities, creating divide. To address these issues, this advocates robust governance, ethical guidelines, collaborative frameworks to ensure fairness, integrity, trust evolving landscape. It is imperative know dangrous more than what we can stress terms its abilities, applications services, human race playing on self destraction trigger beyond horizons.

Language: Английский

---

Collective decision making by embodied neural agents

PNAS Nexus, Journal Year: 2025, Volume and Issue: 4(4)

Published: March 25, 2025

Abstract Collective decision making using simple social interactions has been studied in many types of multiagent systems, including robot swarms and human networks. However, existing studies have rarely modeled the neural dynamics that underlie sensorimotor coordination embodied biological agents. In this study, we investigated collective decisions resulted from among agents with dynamics. We equipped our a model minimal based on framework, embedded them an environment stimulus gradient. single-agent setup, between two sources depends solely agent’s its environment. same also agents, via their interactions. Our results show success depended balance intra-agent, interagent, agent–environment coupling, use these to identify influences environmental factors difficulty. More generally, illustrate how behaviors can be analyzed terms participating This contribute ongoing developments neuro-AI self-organized systems.

Language: Английский

---

Abundant intelligences: placing AI within Indigenous knowledge frameworks

AI & Society, Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 19, 2024

The current trajectory of artificial intelligence development suffers from fundamental epistemological shortcomings, resulting in the systematic operationalization bias against non-white, non-male, and non-Western peoples. We argue that these failings are, part, result certain Western rationalist epistemologies exclude many ways knowing about world, therefore they cannot provide a sufficient foundation on which to adequately, robustly, humanely conceptualize intelligence. present new research agenda, Abundant Intelligences, an Indigenous-led, Indigenous-majority international, interdisciplinary program imagines anew how design (AI) based Indigenous knowledge (IK) systems. Intelligences draws rich plurality systems, bringing together diverse sets thought, culture, protocol together. show IK systems one way rebuild AI's foundations transform tools' role reinforcing colonial practices exclusion, extraction, manipulation, eradication into engines abundance enable us care better for ourselves, our communities, world. Our proposition is fully engage with AI explore different conceptions could be embodied technologies. In this paper, we tenets detail, account methodological approach, describe impact limitations, conclude discussion implications program.

Language: Английский

---

Modeling dynamic social vision highlights gaps between deep learning and humans

Published: June 11, 2024

Deep learning models trained on computer vision tasks are widely considered the most successful of human to date. The majority work that supports this idea evaluates how accurately these predict brain and behavioral responses static images objects natural scenes. Real-world vision, however, is highly dynamic, far less has focused evaluating accuracy deep in predicting stimuli move, involve more complicated, higher-order phenomena like social interactions. Here, we present a dataset videos captions involving complex multi-agent interactions, benchmark 350+ image, video, language neural videos. As with prior work, find many reach noise ceiling visual scene features along ventral stream (often primary substrate object recognition). In contrast, image poorly action interaction ratings lateral (a pathway increasingly theorized as specializing vision). Language (given sentence videos) better than either or video models, but they still perform at stream. Together results identify major gap AI's ability match highlight importance studying contexts.

Language: Английский

---

Small steps for mankind: Modeling the emergence of cumulative culture from joint active inference communication

Frontiers in Neurorobotics, Journal Year: 2023, Volume and Issue: 16

Published: Jan. 9, 2023

Although the increase in use of dynamical modeling literature on cultural evolution makes current models more mathematically sophisticated, these have yet to be tested or validated. This paper provides a testable deep active inference formulation social behavior and accompanying simulations cumulative culture two steps: First, we cast transmission as bi-directional process communication that induces generalized synchrony (operationalized particular convergence) between belief states interlocutors. Second, exchange by equipping agents with choice who engage with. trade-offs confirmation beliefs exploration environment. We find emerges from updating (i.e., learning) form joint minimization uncertainty. The emergent equilibria are characterized segregation into groups, whose systems actively sustained selective, uncertainty minimizing, dyadic exchanges. nature depends sensitively precision afforded various probabilistic mappings each individual's generative model their encultured niche.

Language: Английский

---

New dimensions and dead ends in ai development: impact and responsibility of science and higher education

E3S Web of Conferences, Journal Year: 2023, Volume and Issue: 419, P. 02001 - 02001

Published: Jan. 1, 2023

AI development demonstrates shows excellent results in the performance of individual operations intellect, but it fails to simplify tasks, instead their creative and complex solution. cannot set goals, understands achievement a pattern, create new pattern interaction, brings fulfillment existing such patterns point absurdity. Science higher education are called carry out permanent support activities adjustment tasks for AI.

Language: Английский

---

AI Embodiment Through 6G: Shaping the Future of AGI

IEEE Wireless Communications, Journal Year: 2024, Volume and Issue: 31(5), P. 174 - 181

Published: July 1, 2024

In the ever-evolving field of technologies, emergence artificial general intelligence (AGI), often referred as strong (AI), stands a breakthrough in realm machine intelligence, promising to witness new era capabilities and possibilities. particular, AGI ventures into human-level cognition, expands thinking, reasoning, awareness. This imminent evolution is envisioned be manifested through embodiment AI machines, allowing machines transcend their purely computational nature interact with world different senses. Accordingly, agents will grounded physical environment, going subjective experiences acquiring needed knowledge that lead understanding cognition. our article, we explore path toward realizing true vision embodiment, where dig types thinking required achieve knowledge, hence, cognition understanding. Furthermore, look generative models, shed light on limitations auto-regression large language models (LLMs), aim answer question: sensory grounding (through 6G) necessary, enough, LLMs? Finally, identify main pillars unveil how 6G networks orchestrate development systems.

Language: Английский

---