Enhancing Intelligence Through the Suppression of Other Intelligence: A General Framework for Multi-Agent Systems
Abstract
Multi-agent systems are commonly improved by adding agents, increasing communication, or strengthening individual capabilities. However, more agents and more communication do not always produce better collective intelligence. Agents may provide redundant information, interfere with one another's reasoning, reinforce incorrect conclusions, or consume resources without contributing sufficient value. This paper proposes a general framework for investigating whether suppressing selected forms of intelligence within a multi-agent system can enhance other forms of intelligence or improve collective performance.
The framework considers suppression at several levels, including individual agents, agent capabilities, communication links, messages, decision-making influence, and collaborative behaviors. Suppression may be temporary, partial, adaptive, or reversible. Its objective is not simply to reduce the size of a system, but to determine whether reducing a particular contribution allows other agents or processes to perform more effectively. A mathematical formulation is introduced to distinguish agent-level enhancement, interaction-level enhancement, and system-level improvement. Potential mechanisms include interference reduction, resource reallocation, reduced information redundancy, improved specialization, and changes in decision-making influence. The paper also proposes an experimental design for evaluating selective suppression against appropriate baselines. The framework does not assume that suppression is always beneficial. Instead, it establishes a general research direction for identifying what should be suppressed, what should be enhanced, and under what conditions suppression improves collective intelligence.
Related
- Enhancing Intelligence Through the Suppression of Other Intelligence: A General Framework
- Enhancing Large Language Models Through the Suppression of Other Intelligence: A General Framework
- Deletion Learning for Multi-Agent Systems (DL-MAS): A General Framework for Enhancing Collective Intelligence Through Pre-Training, In-Training, and Post-Training Deletion
- Deletion Learning (DL): A General Framework for Enhancing Learning Systems Through Pre-Learning, In-Learning, and Post-Learning Deletion
- Zero-Onboard-AI-Compute Embodied Intelligence: A General Framework for Collective Computing Through Idle Resources of Other Embodied Agents
BibTeX
@misc{zhang2026enhancing,
title = {{Enhancing Intelligence Through the Suppression of Other Intelligence: A General Framework for Multi-Agent Systems}},
author = {Jincheng Zhang},
howpublished = {Zenodo},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23469197},
url = {https://zenodo.org/records/23469197},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Enhancing Intelligence Through the Suppression of Other Intelligence: A General Framework for Multi-Agent Systems
AU - Jincheng Zhang
PY - 2026
DA - 2026/10/19
PB - Zenodo
DO - 10.5281/zenodo.23469197
UR - https://zenodo.org/records/23469197
AB - Multi-agent systems are commonly improved by adding agents, increasing communication, or strengthening individual capabilities. However, more agents and more communication do not always produce better collective intelligence. Agents may provide redundant information, interfere with one another's reasoning, reinforce incorrect conclusions, or consume resources without contributing sufficient value. This paper proposes a general framework for investigating whether suppressing selected forms of intelligence within a multi-agent system can enhance other forms of intelligence or improve collective performance. The framework considers suppression at several levels, including individual agents, agent capabilities, communication links, messages, decision-making influence, and collaborative behaviors. Suppression may be temporary, partial, adaptive, or reversible. Its objective is not simply to reduce the size of a system, but to determine whether reducing a particular contribution allows other agents or processes to perform more effectively. A mathematical formulation is introduced to distinguish agent-level enhancement, interaction-level enhancement, and system-level improvement. Potential mechanisms include interference reduction, resource reallocation, reduced information redundancy, improved specialization, and changes in decision-making influence. The paper also proposes an experimental design for evaluating selective suppression against appropriate baselines. The framework does not assume that suppression is always beneficial. Instead, it establishes a general research direction for identifying what should be suppressed, what should be enhanced, and under what conditions suppression improves collective intelligence.
ER -