Belief sharing: a blessing or a curse

• URL: http://arxiv.org/abs/2407.02465v1
• Date: Tue, 2 Jul 2024 17:46:42 GMT
• Title: Belief sharing: a blessing or a curse
• Authors: Ozan Catal, Toon Van de Maele, Riddhi J. Pitliya, Mahault Albarracin, Candice Pattisapu, Tim Verbelen,
• Abstract summary: Communication can be cast as sharing beliefs between free-energy minimizing agents. We demonstrate that naively sharing posterior beliefs can give rise to the negative social dynamics of echo chambers and self-doubt.
• Score: 3.2614942160776823
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: When collaborating with multiple parties, communicating relevant information is of utmost importance to efficiently completing the tasks at hand. Under active inference, communication can be cast as sharing beliefs between free-energy minimizing agents, where one agent's beliefs get transformed into an observation modality for the other. However, the best approach for transforming beliefs into observations remains an open question. In this paper, we demonstrate that naively sharing posterior beliefs can give rise to the negative social dynamics of echo chambers and self-doubt. We propose an alternate belief sharing strategy which mitigates these issues.

Related papers

• Neural embedding of beliefs reveals the role of relative dissonance in human decision-making [6.558951808581431]
  We propose a method for extracting nuanced relations between thousands of beliefs by leveraging large-scale user participation data from an online debate platform. This belief embedding space effectively encapsulates the interconnectedness of diverse beliefs as well as polarization across various social issues. We find that the relative distance between one existing beliefs and new beliefs can serve as a quantitative estimate of cognitive dissonance.
  arXiv  Detail & Related papers  (2024-08-13T23:58:45Z)
• Grounding Language about Belief in a Bayesian Theory-of-Mind [5.058204320571824]
  We take a step towards an answer by grounding the semantics of belief statements in a Bayesian theory-of-mind. By modeling how humans jointly infer coherent sets of goals, beliefs, and plans, our framework provides a conceptual role semantics for belief. We evaluate this framework by studying how humans attribute goals and beliefs while watching an agent solve a doors-and-keys gridworld puzzle.
  arXiv  Detail & Related papers  (2024-02-16T02:47:09Z)
• Leveraging Contextual Counterfactuals Toward Belief Calibration [1.418033127602866]
  meta-alignment problem is that human beliefs are diverse and not aligned across populations. In high regret situations, we observe that contextual counterfactuals and recourse costs are important in updating a decision maker's beliefs and the strengths to which such beliefs are held. We introduce the belief calibration cycle' framework to more holistically calibrate this diversity of beliefs with context-driven counterfactual reasoning.
  arXiv  Detail & Related papers  (2023-07-13T01:22:18Z)
• MindDial: Belief Dynamics Tracking with Theory-of-Mind Modeling for Situated Neural Dialogue Generation [62.44907105496227]
  MindDial is a novel conversational framework that can generate situated free-form responses with theory-of-mind modeling. We introduce an explicit mind module that can track the speaker's belief and the speaker's prediction of the listener's belief. Our framework is applied to both prompting and fine-tuning-based models, and is evaluated across scenarios involving both common ground alignment and negotiation.
  arXiv  Detail & Related papers  (2023-06-27T07:24:32Z)
• Flexible social inference facilitates targeted social learning when rewards are not observable [58.762004496858836]
  Groups coordinate more effectively when individuals are able to learn from others' successes. We suggest that social inference capacities may help bridge this gap, allowing individuals to update their beliefs about others' underlying knowledge and success from observable trajectories of behavior.
  arXiv  Detail & Related papers  (2022-12-01T21:04:03Z)
• The language of opinion change on social media under the lens of communicative action [22.70882875813238]
  We identify key ingredients to opinion change by looking at more than 46k posts and more than 3.5M comments on Reddit's r/ChangeMyView. Among the various social dimensions, the ones that are most likely to produce an opinion change are knowledge, similarity, and trust. In line with theories of constructive conflict, our findings show that voicing conflict in the context of a structured public debate can promote integration.
  arXiv  Detail & Related papers  (2022-10-31T11:35:06Z)
• Do Language Models Have Beliefs? Methods for Detecting, Updating, and Visualizing Model Beliefs [76.6325846350907]
  Dennett (1995) famously argues that even thermostats have beliefs, on the view that a belief is simply an informational state decoupled from any motivational state. In this paper, we discuss approaches to detecting when models have beliefs about the world, and we improve on methods for updating model beliefs to be more truthful.
  arXiv  Detail & Related papers  (2021-11-26T18:33:59Z)
• Shades of confusion: Lexical uncertainty modulates ad hoc coordination in an interactive communication task [8.17947290421835]
  We propose a communication task based on color-concept associations. In Experiment 1, we establish several key properties of the mental representations of these expectations. In Experiment 2, we examine the downstream consequences of these representations for communication.
  arXiv  Detail & Related papers  (2021-05-13T20:42:28Z)
• Learning Triadic Belief Dynamics in Nonverbal Communication from Videos [81.42305032083716]
  Nonverbal communication can convey rich social information among agents. In this paper, we incorporate different nonverbal communication cues to represent, model, learn, and infer agents' mental states.
  arXiv  Detail & Related papers  (2021-04-07T00:52:04Z)
• Causal Understanding of Fake News Dissemination on Social Media [50.4854427067898]
  We argue that it is critical to understand what user attributes potentially cause users to share fake news. In fake news dissemination, confounders can be characterized by fake news sharing behavior that inherently relates to user attributes and online activities. We propose a principled approach to alleviating selection bias in fake news dissemination.
  arXiv  Detail & Related papers  (2020-10-20T19:37:04Z)
• Information Consumption and Social Response in a Segregated Environment: the Case of Gab [74.5095691235917]
  This work provides a characterization of the interaction patterns within Gab around the COVID-19 topic. We find that there are no strong statistical differences in the social response to questionable and reliable content. Our results provide insights toward the understanding of coordinated inauthentic behavior and on the early-warning of information operation.
  arXiv  Detail & Related papers  (2020-06-03T11:34:25Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.