Pith Number

pith:UZWPQKON

pith:2026:UZWPQKONS7CKVFUR4EEUBX2TGW

not attested not anchored not stored refs resolved

Reinforced Collaboration in Multi-Agent Flow Networks

Yangkai Ding, Yuang Liu, Zheng Wang

MANGO improves multi-agent LLM collaboration by building flow networks from successful workflows and optimizing them with reinforcement learning and textual gradients.

arxiv:2605.12943 v1 · 2026-05-13 · cs.LG

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{UZWPQKONS7CKVFUR4EEUBX2TGW}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

MANGO achieves up to 12.8% performance improvement over state-of-the-art baselines, enhances efficiency by 47.4%, and generalizes effectively to unseen domains.

C2weakest assumption

That flow networks built from past successful workflows, when optimized by RL and textual gradients, will reliably reduce error propagation and generalize without the optimization itself introducing new failure modes or overfitting to the training workflows.

C3one line summary

MANGO optimizes multi-agent LLM workflows via flow networks, RL, and textual gradients, delivering up to 12.8% higher performance and 47.4% better efficiency while generalizing to new domains.

References

93 extracted · 93 resolved · 12 Pith anchors

[1] Program Synthesis with Large Language Models 2021 · arXiv:2108.07732

[2] Harrison Chase. Langchain.https://github.com/langchain-ai/langchain, 2022 2022

[3] Evaluating Large Language Models Trained on Code 2021 · arXiv:2107.03374

[4] Training Verifiers to Solve Math Word Problems 2021 · arXiv:2110.14168

[5] Multi-agent collaboration via evolving orchestration.NeurIPS, 2025 2025

Receipt and verification

First computed	2026-05-18T03:09:09.585294Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

a66cf829cd97c4aa9691e10940df5335916126c53e514fe1b144ae859fd0f13c

Aliases

arxiv: 2605.12943 · arxiv_version: 2605.12943v1 · doi: 10.48550/arxiv.2605.12943 · pith_short_12: UZWPQKONS7CK · pith_short_16: UZWPQKONS7CKVFUR · pith_short_8: UZWPQKON

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/UZWPQKONS7CKVFUR4EEUBX2TGW \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: a66cf829cd97c4aa9691e10940df5335916126c53e514fe1b144ae859fd0f13c

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "7f934e99869c6f5bd18c60b1be0d295b7c350bb9e91edbabac8fd741b264c9cd",
    "cross_cats_sorted": [],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-05-13T03:26:07Z",
    "title_canon_sha256": "eb431d9231f1bf457fa4101afba56973f2ffaea3b27f98cf01e034650b30131e"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.12943",
    "kind": "arxiv",
    "version": 1
  }
}