REVIEW 35 cited by
Scaling Large Language Model-based Multi-Agent Collaboration
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
Recent breakthroughs in large language model-driven autonomous agents have revealed that multi-agent collaboration often surpasses each individual through collective reasoning. Inspired by the neural scaling law--increasing neurons enhances performance, this study explores whether the continuous addition of collaborative agents can yield similar benefits. Technically, we utilize directed acyclic graphs to organize agents into a multi-agent collaboration network (MacNet), upon which their interactive reasoning is topologically orchestrated for autonomous task solving. Extensive evaluations reveal that it effectively supports collaboration among over a thousand agents, with irregular topologies outperforming regular ones. We also identify a collaborative scaling law--the overall performance follows a logistic growth pattern as agents scale, with collaborative emergence occurring earlier than traditional neural emergence. We speculate this may be because scaling agents catalyzes their multidimensional considerations during interactive reflection and refinement, thereby producing more comprehensive artifacts. The code is available at https://github.com/OpenBMB/ChatDev/tree/macnet.
Forward citations
Cited by 35 Pith papers
-
MoRSE: Task-Oriented Multi-Agent System with Mixture of Role-Subtask Experts
MoRSE trains role- and subtask-specific LoRA experts with a prototype router and hierarchical GRPO, improving LLM code-generation benchmarks and held-out task generalization.
-
MasFACT: Continual Multi-Agent Topology Learning via Geometry-Aware Posterior Transfer
MasFACT transfers historical topology priors across tasks via Fused Gromov-Wasserstein optimal transport and PAC-Bayes conservative adaptation to reduce topology forgetting in continual multi-agent settings.
-
LLMs Can Simulate Standardized Patients via Agent Coevolution
EvoPatient uses unsupervised coevolution of patient and doctor LLM agents to build a retrieval library that makes simulated patients more faithful, robust, and preferred by human experts than reasoning-only baselines.
-
Width, Memory, and Delay: A Resource Accounting for the Limits of Flat Multi-Agent Systems
A flat swarm with matched internal disturbance models can match or beat a two-loop hierarchy at equal per-agent memory; width cannot replace memory, and delay sets an irreducible floor.
-
When Agents Go Rogue: Activation-Based Detection of Malicious Behaviors in Multi-Agent Systems
Activation-space divergence detects and corrects compromised LLM agents in multi-agent systems without interaction graphs or synchronized rounds, outperforming graph baselines especially under async stealthy attacks.
-
DECIS: Dual-Evidence Corrective Verification for Interpretable Strabismus Diagnostic Decision-Making
Structured evidence-constrained verification with multi-agent reasoning improves strabismus subtype diagnosis from 72.0% to 91.3% weighted F1 and increases clinical report reliability.
-
Accelerating Heterogeneous Agent Collaboration in Dynamic Edge Networks
PRADA distills a process reward model into an edge screening policy and uses a threshold-based server scheduler to retain most LLM reasoning accuracy while sharply cutting multi-user latency.
-
Effective Strategies for Asynchronous Software Engineering Agents
CAID, a manager-driven multi-agent system using git worktrees, commits, and merges, improves long-horizon SWE success by roughly 14–27 absolute points over single-agent baselines.
-
Optimal-Agent-Selection: State-Aware Routing Framework for Efficient Multi-Agent Collaboration
A state-aware contrastive router that selects the most relevant agent at each step improves multi-agent LLM accuracy by up to 23.8% while using a fraction of the tokens of fixed-pipeline baselines.
-
SWE-Debate: Competitive Multi-Agent Debate for Software Issue Resolution
A competitive multi-agent debate over graph-derived fault propagation chains lifts automated issue resolution on SWE-bench Verified to 41.4% pass@1.
-
Think Like an Engineer: A Neuro-Symbolic Collaboration Agent for Generative Software Requirements Elicitation and Self-Review
RequireCEG combines large language models with causal-effect graphs to elicit and self-review Gherkin requirements from natural language narratives, reporting improved quality, diversity, and consistency over baselines.
-
AgentsNet: Coordination and Collaborative Reasoning in Multi-Agent LLMs
A benchmark built from five distributed computing problems shows that frontier LLM agent networks solve small coordination tasks but break down as the network scales to 100 agents.
-
A Theory of Inference Compute Scaling: Reasoning through Directed Stochastic Skill Search
A skill-graph random-walk model gives closed-form accuracy-versus-compute formulas for four reasoning strategies and connects them to training scaling.
-
G-Memory: Tracing Hierarchical Memory for Multi-Agent Systems
G-Memory stores past multi-agent teamwork in a three-tier graph and retrieves it to boost performance on five benchmarks.
-
Adaptive Graph Pruning for Multi-Agent Communication
AGP trains a graph neural network to jointly decide which agents to keep and how strongly they should communicate, and reports state-of-the-art average accuracy across six LLM benchmarks with large token savings.
-
Stop Overvaluing Multi-Agent Debate -- We Must Rethink Evaluation and Embrace Model Heterogeneity
Multi-agent debate mostly underperforms simple chain-of-thought baselines when tested broadly, while randomly mixing different models into the debate reliably improves performance.
-
SYMPHONY: Improving Memory Management for LLM Inference Workloads
SYMPHONY uses advisory requests that signal upcoming multi-turn LLM turns to prefetch KV caches, enabling request-level load balancing and up to 8x more served requests in ShareGPT-based tests.
-
Is Inter-Seed Cross-Play Enough? Evaluating the Robustness of Zero-Shot Coordination Algorithms to Implementation Details
For Other-Play in Yokai, agents trained with different implementation details coordinate across implementations about as well as across seeds, supporting inter-seed cross-play as a proxy for cross-implementation evaluation.
-
Scaling LLM-Driven Multi-Agent Systems: Design Principles and Architectural Scalability Analysis
Architecturally scaling LLM multi-agent systems raises accuracy at near-linear cost only above a model-capability threshold, peaks at intermediate complexity, and never fixes poor run-to-run consistency.
-
Transition from Statistical to Hardware-Limited Scaling in Photonic Quantum State Reconstruction
Classical shadow tomography on integrated photonics shows a sharp transition from statistical O(M^{-1/2}) error scaling to a hardware-limited floor set by unitary spectral distortions.
-
SERM: Self-Evolving Relevance Model with Agent-Driven Learning from Massive Query Streams
SERM deploys multi-agent sample mining and two-level label agreement to enable iterative self-evolution of relevance models on industrial query streams, yielding performance gains in offline and online tests.
-
An Auditable Agent Platform For Automated Molecular Optimisation
A hierarchical multi-agent LLM platform with recorded provenance improved average predicted binding affinity for AKT1 by 31%, while single-agent runs favored drug-likeness.
-
A Framework for Analyzing Abnormal Emergence in Service Ecosystems Through LLM-based Agent Intention Mining
EAMI extracts agent intentions from LLM thought traces, clusters them by meaning, and builds temporal diagrams that trace how new intentions spread through a simulated ecosystem.
-
Rethinking the Illusion of Thinking
Reasoning models' Towers of Hanoi failures persist under stepwise prompting, while River Crossing failures mostly vanish when tests are restricted to solvable configurations.
-
An Adversary-Resistant Multi-Agent LLM System via Credibility Scoring
A credibility-scoring framework for multi-agent LLM systems, learning agent trustworthiness on the fly and weighting outputs accordingly, improves accuracy under adversarial conditions in some benchmarks.
-
MermaidFlow: Redefining Agentic Workflow Generation via Safety-Constrained Evolutionary Programming
Encoding LLM agent workflows as statically checked Mermaid graphs and evolving them with constrained mutations improves solve rates on GSM8K, MATH, HumanEval, and MBPP.
-
BetaWeb: Towards a Blockchain-enabled Trustworthy Agentic Web
BetaWeb promises a blockchain-enabled trustworthy agentic web, but the submitted manuscript body is a different mining-robot paper, leaving the proposal without supporting evidence.
-
How Far Are AI Scientists from Changing the World?
This survey proposes a four-level capability framework for AI Scientist systems and, using an AI reviewer, finds that current systems produce papers rated well below normal scientific standards.
-
Enhancing LLM Code Generation: A Systematic Evaluation of Multi-Agent Collaboration and Runtime Debugging for Improved Accuracy, Reliability, and Latency
Chaining a simple two-agent coding workflow with execution-based debugging improves code-generation accuracy by about 0.7 percent over debugging alone, but the improvement is not statistically robust.
-
OMoE: Diversifying Mixture of Low-Rank Adaptation by Orthogonal Finetuning
OMoE applies Gram-Schmidt orthogonalization to expert representations in LoRA-based mixture-of-experts, improving fine-tuning performance with only two experts.
-
Flow: Modularized Agentic Workflow Automation
Flow represents a task as a dependency graph of subtasks and lets LLM agents redraw that graph during execution, reporting better success rates than three baselines on three coding tasks.
-
A Survey on LLM-based Multi-Agent System: Recent Advances and New Frontiers in Application
This survey organizes recent LLM-based multi-agent research into task-solving, simulation, and agent-evaluation applications, and identifies efficiency and evaluation gaps as key open problems.
-
From Individual to Society: A Survey on Social Simulation Driven by Large Language Model-based Agents
A structured survey that categorizes LLM-based social simulation into individual, scenario, and society simulation, with associated methods, benchmarks, and observed trends.
-
Toward Efficient Agents: Memory, Tool learning, and Planning
A survey that organizes efficiency techniques for LLM agents into memory, tool learning, and planning, and consolidates benchmarks and metrics for measuring cost-performance trade-offs.
-
SIGMA: Search-Augmented On-Demand Knowledge Integration for Agentic Mathematical Reasoning
SIGMA uses four specialized retrieval agents plus a moderator to improve math reasoning, reporting up to 7.4 absolute MATH500 points over Search-o1 at 1.5B scale.
Discussion (0). Continue with ORCID to comment.