Research map / Bandits and online learning / Online convex optimization and games
Last-iterate convergence in games
Prove convergence of uncoupled no-regret dynamics in zero-sum, monotone, and potential games.
65 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Online convex optimization and games, keyword cluster “game · equilibrium · convergence”
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.
15 papers in 2023–24 and 19 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 6.6% to 6.9%.
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
- 2017 · Learning with Bandit Feedback in Potential Games (NeurIPS 2017 · 53 citations)
- 2018 · Bandit Learning in Concave N-Person Games (NeurIPS 2018 · 28 citations)
- 2019 · Fast and Furious Learning in Zero-Sum Games: Vanishing Regret with Non-Vanishing Step Sizes (NeurIPS 2019 · 17 citations)
- 2020 · Tight last-iterate convergence rates for no-regret learning in multi-player games (NeurIPS 2020 · 16 citations)
- 2021 · Faster Game Solving via Predictive Blackwell Approachability: Connecting Regret Matching and Mirror Descent (AAAI 2021 · 23 citations)
- 2022 · Learning to Identify Top Elo Ratings: A Dueling Bandits Approach (AAAI 2022 · 10 citations)
- 2023 · Logarithmic-Regret Quantum Learning Algorithms for Zero-Sum Games (NeurIPS 2023 · 3 citations)
- 2024 · No-regret Learning in Harmonic Games: Extrapolation in the Face of Conflicting Interests (NeurIPS 2024 · 1 citation)
- 2025 · The Harder Path: Last Iterate Convergence for Uncoupled Learning in Zero-Sum Games with Bandit Feedback (ICML 2025 · 0 citations)
- 2026 · Achieving Logarithmic Regret in KL-Regularized Zero-Sum Markov Games (ICML 2026 · 0 citations)
Earliest papers here
- Learning with Bandit Feedback in Potential Games (NeurIPS 2017 · 53 citations)
- Bandit Learning in Concave N-Person Games (NeurIPS 2018 · 28 citations)
- Multiplicative Weights Updates with Constant Step-Size in Graphical Constant-Sum Games (NeurIPS 2018 · 7 citations)
What has been tried
42 papers here report 239 measured changes (a component added, removed or swapped, compared with a baseline): helped 67%, no clear effect 5%, hurt 21%, mixed 7%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Set optimism weight n>m rather than n≤m | Matching Pennies, m=10 · social regret | O(√T) -> O(1) | Helped | Learning from Delayed Feedback in Games via Extra Prediction (NeurIPS 2025 · 0 citations) |
| use Smooth PTBP instead of PTBP | Kuhn poker, Leduc poker, Liar's Dice, Goofspiel, Battleship · convergence to Nash equilibrium | Hurt | Extensive-Form Game Solving via Blackwell Approachability on Treeplexes (NeurIPS 2024 · 0 citations) | |
| use virtual transition-weighted negentropy instead of vanilla negentropy | IIEFGs · last-iterate NE gap convergence rate | eO((X + Y)[(XA + YB)^{1/2} + (X + Y)^{1/4}H]k^{-1/8}) vs. eO((XA + YB)Hk^{-1/8}); the paper says each rate is superior i | Mixed | Learning Imperfect Information Extensive-form Games with Last-iterate Convergence under Bandit Feedback (ICML 2025) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 34 papers; from OpenAlex and the papers' affiliations).
- Yale University · 3 papers
- Singapore University of Technology and Design · 3 papers
- Columbia University · 3 papers
- Princeton University · 2 papers
- Meta · 2 papers
- Tsinghua University · 2 papers
- OMRON SINIC X · 2 papers
- Google DeepMind · 2 papers
- Shanghai Jiao Tong University · 2 papers
- The Chinese University of Hong Kong · 2 papers
Use it
Download BibTeX of 40 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
- Fast no-regret dynamics in games (37 papers)
- Counterfactual regret minimization (15 papers)
Elsewhere in the topic
- Online learning with feedback graphs (155 papers)
- Best-of-both-worlds bandit algorithms (125 papers)
- Learning for pricing and auctions (101 papers)
- Constrained and unconstrained OCO algorithms (85 papers)
- Online decision-making with feedback structure (81 papers)
- Online allocation with constraints (77 papers)
- Learning-augmented online algorithms (73 papers)
- Dynamic and adaptive regret (57 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).