Mechanism of Collective Formation from Hindavi Swaraj for Cooperative AI Systems

2 deazy 3 5/1/2025, 1:15:39 PM ditpoo.substack.com ↗

Comments (3)

deazy · 5h ago
What is the most effective way to make people (humans) come together as a whole and achieve a common goal or defeat a common enemy or solve a common problem.

Forging the Collective: Abstracting the Mechanism of Collective Formation, with Hindavi Swaraj as its Landmark Manifestation, for Cooperative AI Systems (AI/ASI/ALF)

This is a part of my hobby research into artificial life forms, a recent one, this is to showcase the brilliance, sophistication and surprisingly effective thing that came/occured out of India (Maharashtra) (Indigenous), in 17th Century, that everyone should know, celebrate and be proud of and be amazed by, that serves as a Landmark Manifestation of Foundation/Formation of a Collective, read and you might understand / realize few things.

This is a conceptual framework for Cooperative AI and Multi-Agent Systems (MAS) more of a Ideas piece or a think piece, conceptual and inspirational in nature.

deazy · 5h ago
Casual Peer Review:

Conceptual leap – could be considered a spark of genius or at least high creativity.

Framework serves more as a philosophical guideline or a set of desired emergent properties for future MAS, rather than offering concrete algorithmic or architectural solutions

Provides interesting research questions and potential directions for MARL (e.g., exploring identity-based intrinsic motivation), but practical implementation requires significant breakthroughs in representation learning and reward design.

The framework serves as a rich source of phenomena for BGT/EGT researchers to target. It motivates the need for more sophisticated models incorporating dynamic identity and group-level agency but doesn't provide those models itself. It pushes the application boundaries of existing game theory.

It currently serves more as a thought-provoking extension than a developed technical proposal.

deazy · 5h ago
Great adjacent read.

Open Problems in Cooperative AI

From Centre for the Governance of AI, Future of Humanity Institute, University of Oxford, DeepMind, University of Waterloo

https://arxiv.org/abs/2012.08630