Research map / Bandits and online learning / Multi-armed bandits
Multi-agent bandits
Coordinate multiple or federated agents in decentralized bandit learning.
42 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Multi-armed bandits, keyword cluster “multi armed · armed bandit · stochastic”
State History Tried Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
17 papers in 2023–24 and 10 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 16.0% to 8.0%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Multi-Player Bandits – a Musical Chairs Approach (ICML 2016 · 3 citations)
- 2018 · Distributed Multi-Player Bandits - a Game of Thrones Approach (NeurIPS 2018 · 11 citations)
- 2019 · Decentralized Cooperative Stochastic Bandits (NeurIPS 2019 · 67 citations)
- 2020 · Distributed Bandit Learning: Near-Optimal Regret with Efficient Communication (ICLR 2020 · 7 citations)
- 2021 · Federated Multi-Armed Bandits (AAAI 2021 · 64 citations)
- 2022 · Matching in Multi-arm Bandit with Collision (NeurIPS 2022 · 0 citations)
- 2023 · An Efficient Algorithm for Fair Multi-Agent Multi-Armed Bandit with Low Regret (AAAI 2023 · 3 citations)
- 2024 · Decentralized Scheduling with QoS Constraints: Achieving O(1) QoS Regret of Multi-Player Bandits (AAAI 2024 · 4 citations)
- 2025 · Heterogeneous Multi-Agent Bandits with Parsimonious Hints (AAAI 2025 · 1 citation)
- 2026 · Robust Decentralized Multi-armed Bandits: From Corruption-Resilience to Byzantine-Resilience (AAAI 2026 · 1 citation)
Earliest papers here
- Multi-Player Bandits – a Musical Chairs Approach (ICML 2016 · 3 citations)
- Distributed Multi-Player Bandits - a Game of Thrones Approach (NeurIPS 2018 · 11 citations)
- Partial Truthfulness in Minimal Peer Prediction Mechanisms With Limited Knowledge (AAAI 2018 · 1 citation)
What has been tried
36 papers here report 195 measured changes (a component added, removed or swapped, compared with a baseline): helped 67%, no clear effect 6%, hurt 18%, mixed 10%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use A2deQoS without prior knowledge of ∆ and omit the signaling procedure | MP-MAB · QoS regret | O(1) -> O(log T) | Mixed | Decentralized Scheduling with QoS Constraints: Achieving O(1) QoS Regret of Multi-Player Bandits (AAAI 2024 · 4 citations) |
| Reduce ε from 1 (MATS) to 0.1 | Bernoulli 0101-Chain and Poisson 0101-Chain · cumulative regret | Helped | Finite-Time Frequentist Regret Bounds of Multi-Agent Thompson Sampling on Sparse Hypergraphs (AAAI 2024 · 2 citations) | |
| remove projection from hint inquiries | HMA2B centralized · hint complexity | O\left(\frac{MK \log T}{\Delta^{\mathrm{kl}}}\right) -> O\left(\frac{M^2K \log T}{\Delta^{\mathrm{kl}}}\right) | Hurt | Heterogeneous Multi-Agent Bandits with Parsimonious Hints (AAAI 2025 · 1 citation) |
| Use random action choices instead of round-robin selection | Cooperative stochastic MAB · individual regret | Mixed | Individual Regret in Cooperative Stochastic Multi-Armed Bandits (NeurIPS 2025 · 0 citations) | |
| Use Bernstein-type confidence intervals instead of Hoeffding-type confidence intervals | Stochastic N-agent K-armed bandits, bandit feedback · expected regret | \otil(K^{\frac{1}{2N}}T^{\frac{2N-1}{2N}}) -> \otil(K^{\frac{2}{N}}T^{\frac{N-1}{N}}+K) | Helped | No-Regret Learning for Fair Multi-Agent Social Welfare Optimization (NeurIPS 2024 · 0 citations) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 27 papers; from OpenAlex and the papers' affiliations).
- University of Massachusetts at Amherst · 3 papers
- Tsinghua University · 2 papers
- National University of Singapore · 2 papers
- University of Massachusetts Amherst · 2 papers
- Nanjing University · 2 papers
- University of Milan · 2 papers
Use it
Download BibTeX of 34 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Non-stationary and structured-reward bandits (140 papers)
- Optimal stochastic bandit algorithms (124 papers)
- Bandit formulations for ML problems (42 papers)
Elsewhere in the topic
- Best-arm identification in linear bandits (57 papers)
- Pure exploration algorithms (57 papers)
- Fixed-confidence versus fixed-budget identification (35 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).