When Do Multi-Agent Systems Help? An Information Bottleneck Perspective

• URL: http://arxiv.org/abs/2607.16133v1
• Date: Fri, 17 Jul 2026 17:13:44 GMT
• Title: When Do Multi-Agent Systems Help? An Information Bottleneck Perspective
• Abstract summary: We provide an information bottleneck perspective on elucidating the differences between multi-agent systems and single-agent systems.<n>We show that under infinite relay bandwidth, any SAS can be simulated by a MAS that transmits the full upstream context.<n>We formalize this trade-off as an information bottleneck controlled by an effective parameter $$, which captures how the balance shifts with model capability.
• Score: 58.23673862794022
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: LLM powered multi-agent systems (MAS) have emerged as a promising paradigm for complex tasks. However, their advantages over single-agent systems (SAS) remain unclear, with performance varying inconsistently across settings. Here, we provide an information bottleneck perspective on elucidating the differences between MAS and SAS. Specifically, our key observation is that a SAS accumulates its full reasoning trace in one shared context, while a MAS uses isolated local contexts connected by bounded relay messages. We show that, under infinite relay bandwidth, any SAS can be simulated by a MAS that transmits the full upstream context. Thus, the nontrivial advantage of MAS arises under bounded relays, where compression introduces a fundamental trade-off: reducing redundant context can improve efficiency, but may also incur loss of task-relevant information. We formalize this trade-off as an information bottleneck controlled by an effective parameter $β$, which captures how the balance shifts with model capability, and shows that MAS gains arise when context reduction outweighs relay information loss. We conduct 18 controlled experiments across five benchmarks and three model scales to validate our theoretical studies. We observe that MAS consistently helps when relays are near-sufficient, especially for weaker models. In contrast, MAS gains shrink or reverse when relays incur information loss, especially for stronger models that can already extract useful information from redundant context and thus gain little from compression. Our study shows that multi-agent design is fundamentally an information-bottleneck optimization problem. This perspective explains when bounded inter-agent communication helps or hurts.

Related papers

• Large Vision-Language Models Get Lost in Attention [51.851592109135716]
  We propose a unified framework grounded in information theory and geometry to quantify the geometric and entropic nature of residual updates.<n>Applying this unified framework reveals a fundamental functional decoupling: Attention acts as a subspace-language operator focused on reconfiguration, whereas FFNs serve as subspace-language operators driving semantic innovation.
  arXiv  Detail & Related papers  (2026-05-07T04:45:52Z)
• Does RL Expand the Capability Boundary of LLM Agents? A PASS@(k,T) Analysis [16.739617199649615]
  We introduce PASS@(k,T), a metric that separates capability expansion from efficiency improvement.<n>Our main finding is that, contrary to the static-reasoning result, tool-use RL genuinely enlarges the capability boundary.<n>Under matched training data, supervised fine-tuning regresses the boundary on the same compositional tasks, isolating self-directed exploration as the causal factor.
  arXiv  Detail & Related papers  (2026-04-16T11:06:19Z)
• Breaking Contextual Inertia: Reinforcement Learning with Single-Turn Anchors for Stable Multi-Turn Interaction [49.03500737694832]
  We introduce textbfReinforcement textbfLearning with textbfTurn textbfRLSTA, a generalizable training approach designed to stabilize multi-turn interaction.<n>Experiments show that RLSTA significantly outperforms standard fine-tuning and abstention-based methods.
  arXiv  Detail & Related papers  (2026-03-05T04:04:59Z)
• TopoCurate:Modeling Interaction Topology for Tool-Use Agent Training [53.93696896939915]
  Training tool-use agents typically rely on Supervised Fine-Tuning (SFT) on successful trajectories and Reinforcement Learning (RL) on pass-rate-selected tasks.<n>We propose TopoCurate, an interaction-aware framework that projects multi-trial rollouts from the same task into a unified semantic quotient topology.<n>TopoCurate achieves consistent gains of 4.2% (SFT) and 6.9% (RL) over state-of-the-art baselines.
  arXiv  Detail & Related papers  (2026-03-02T10:38:54Z)
• LSTM-MAS: A Long Short-Term Memory Inspired Multi-Agent System for Long-Context Understanding [24.027208865014064]
  Long language models (LLMs) are difficult to process because of the accumulation of errors and the propagation of hallucinations.<n>We design a Multi-Agent System called LSTM-MAS, emulating LSTM's hierarchical information flow and gated memory mechanisms for long-context understanding.<n>Our model achieves improvements of 40.93%, 43.70%,121.57% and 33.12%, on NarrativeQA, Qasper, HotpotQA, and MuSiQue, respectively.
  arXiv  Detail & Related papers  (2026-01-17T05:16:23Z)
• Explainability of Complex AI Models with Correlation Impact Ratio [10.61008729196936]
  Complex AI systems make better predictions but often lack transparency, limiting trustworthiness, interpretability, and safe deployment.<n>We introduce ExCIR (Explainability through Correlation Impact Ratio), a theoretically grounded, simple, and reliable metric for explaining the contribution of input features to model outputs.<n>We demonstrate that ExCIR captures dependencies arising from correlated features through a lightweight single pass formulation.
  arXiv  Detail & Related papers  (2026-01-10T21:56:24Z)
• RMoA: Optimizing Mixture-of-Agents through Diversity Maximization and Residual Compensation [6.364685086217188]
  We propose Residual Mixture-of-Agents (RMoA) to integrate residual connections to optimize efficiency and reliability.<n>RMoA achieves state-of-the-art performance on the benchmarks of across alignment, mathematical reasoning, code generation, and multitasking understanding.
  arXiv  Detail & Related papers  (2025-05-30T10:23:11Z)
• SAMSA: Efficient Transformer for Many Data Modalities [12.7600763629179]
  We propose SAMSA - SAMpling-Self-Attention, a context-aware linear complexity self-attention mechanism. Our mechanism is based on a differentiable sampling without replacement method we discovered. SAMSA achieved competitive or even SOTA results on many benchmarks, while being faster in inference, compared to other very specialized models.
  arXiv  Detail & Related papers  (2024-08-10T00:09:06Z)
• Enhancing data efficiency in reinforcement learning: a novel imagination mechanism based on mesh information propagation [0.3729614006275886]
  We introduce a novel mesh information propagation mechanism, termed the 'Imagination Mechanism (IM)' IM enables information generated by a single sample to be effectively broadcasted to different states across episodes. To promote versatility, we extend the IM to function as a plug-and-play module that can be seamlessly and fluidly integrated into other widely adopted RL algorithms.
  arXiv  Detail & Related papers  (2023-09-25T16:03:08Z)
• Correlation Information Bottleneck: Towards Adapting Pretrained Multimodal Models for Robust Visual Question Answering [63.87200781247364]
  Correlation Information Bottleneck (CIB) seeks a tradeoff between compression and redundancy in representations. We derive a tight theoretical upper bound for the mutual information between multimodal inputs and representations.
  arXiv  Detail & Related papers  (2022-09-14T22:04:10Z)
• Examining and Combating Spurious Features under Distribution Shift [94.31956965507085]
  We define and analyze robust and spurious representations using the information-theoretic concept of minimal sufficient statistics. We prove that even when there is only bias of the input distribution, models can still pick up spurious features from their training data. Inspired by our analysis, we demonstrate that group DRO can fail when groups do not directly account for various spurious correlations.
  arXiv  Detail & Related papers  (2021-06-14T05:39:09Z)
• Understanding Self-supervised Learning with Dual Deep Networks [74.92916579635336]
  We propose a novel framework to understand contrastive self-supervised learning (SSL) methods that employ dual pairs of deep ReLU networks. We prove that in each SGD update of SimCLR with various loss functions, the weights at each layer are updated by a emphcovariance operator. To further study what role the covariance operator plays and which features are learned in such a process, we model data generation and augmentation processes through a emphhierarchical latent tree model (HLTM)
  arXiv  Detail & Related papers  (2020-10-01T17:51:49Z)

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.