A Q-learning algorithm finds Nash equilibrium in two-player stochastic games efficiently.
problem Finding Nash equilibrium in two-player stochastic games with limited samples.
method Feature-based Q-learning algorithm with accelerated techniques for improved sample efficiency.
result The algorithm finds an ε-optimal strategy with sample size linear to the number of features and time/space complexity independent of game dimensions.
Novel algorithms for multi-agent reinforcement learning reduce sample complexity.
problem Efficiently learning Nash equilibria in multi-agent settings.
method Information-Directed Sampling (IDS) principles applied to multi-agent reinforcement learning.
result Sample-efficient algorithms for learning Nash equilibria in various multi-agent settings.
This paper improves sample efficiency for learning equilibria in multi-player games.
problem Sample-efficient learning of equilibria in games with many players.
method Designs algorithms for learning CCE and CE with polynomial sample complexity in the number of players.
result First to show polynomial sample complexity for learning CCE and CE in multi-player games.
New method finds Nash equilibria faster in multiplayer games.
problem Finding Nash equilibria in multi-player games, especially in noisy environments.
method Extra-gradient with player sampling and variance reduction.
result Proves better convergence rate than full extra-gradient for noisy games.
We study the problem of learning classifiers robust to universal adversarial perturbations. While prior work approaches this problem via robust optimization, adversarial training, or input transformation, we instead phrase it as a two-player zero-sum game. In this new formulation, both players simultaneously play the s…
Paper solves discounted stochastic games with near-optimal time and sample complexity.
problem Solving discounted stochastic two-player games with optimal complexity.
method Generalizes Q-learning to two-player strategy computation, overcoming limitations of existing methods.
result Near-optimal ε ε ε -strategy computation with polylogarithmic factors in 1 − γ 1 - γ 1 − γ and ε − 2 ε^{-2} ε − 2 . Optimistic Thompson Sampling reduces regret in unknown multi-player games.
problem Navigating uncertainty in unknown multi-player games with strategic decision-making.
method Introduces Thompson Sampling algorithms that exploit opponents' actions and reward structures.
result Achieves over tenfold improvements in experimental budgets with logarithmic regret bound.
Graphon game model simplifies stochastic interactions among agents.
problem Complex interactions among heterogeneous agents in stochastic games.
method Introduced a discrete-time graphon game formulation with a representative player.
result Existence and uniqueness of graphon equilibrium proven with mild assumptions.
A game-theoretic approach simplifies MBRL design and improves sample efficiency.
problem Designing stable and efficient MBRL algorithms using rich function approximators.
method Develops a game-theoretic framework where MBRL is modeled as a Stackelberg game between policy and model players.
result Proposed algorithms are highly sample efficient and match asymptotic performance of model-free policy gradient.
Fictitious play is a simple and widely studied adaptive heuristic for playing repeated games. It is well known that fictitious play fails to be Hannan consistent. Several variants of fictitious play including regret matching, generalized regret matching and smooth fictitious play, are known to be Hannan consistent. In …
Study analyzes player churn and purchasing behavior in games.
problem Retaining players and predicting churn in video games.
method Deep behavioral analysis and ensemble learning models.
result Discarding certain churners improves prediction models.
Paper optimizes multi-agent learning in Markov games with generative model.
problem Learning Nash or CCE equilibria in multi-agent Markov games.
method Develops \myalg~algorithm and adaptive sampling scheme using FTRL method.
result Minimax-optimal learning of CCE with minimal samples.
Proves exponential sample complexity separations in local differential privacy.
problem Sample complexity in locally private protocols.
method Connection between communication complexity and sample complexity, using specific lower bounds for two problems.
result Exponential separations between differentially private protocols.
Paper develops efficient algorithms for learning rationalizable equilibria in multiplayer games.
problem Learning rationalizable behavior in multiplayer games under bandit feedback.
method New algorithms for finding rationalizable Coarse Correlated Equilibria and Correlated Equilibria with polynomial sample complexity.
result Achieved polynomial sample complexity for learning rationalizable equilibria, improving over existing exponential complexity.
New RL method improves robustness across various conditions.
problem Training robust RL agents for diverse environments.
method Adversarial training with Langevin Dynamics.
result Consistently outperforms existing RL algorithms in generalization.
Pessimistic model-based algorithm finds Nash equilibria in zero-sum Markov games from offline data.
problem Learning Nash equilibria in two-player zero-sum Markov games from limited data.
method Pessimistic model-based algorithm with Bernstein-style lower confidence bounds (VI-LCB-Game).
result Proves sample complexity no larger than C c l i p p e d ⋆ S ( A + B ) ( 1 − γ ) 3 ε 2 \frac{C_{\mathsf{clipped}}^\star S(A+B)}{(1-γ)^3 \varepsilon^2} ( 1 − γ ) 3 ε 2 C clipped ⋆ S ( A + B ) , achieving minimax optimality. New algorithm identifies optimal actions in large reward spaces efficiently.
problem Finding the best action from a large set of options with minimal trials.
method GenTS-Explore algorithm for real-valued combinatorial pure exploration.
result Achieves optimal sample complexity for large action sets.
Paper solves learning imperfect-information games with fewer episodes.
problem Learning imperfect-information extensive-form games from bandit feedback.
method Balanced Online Mirror Descent and Balanced Counterfactual Regret Minimization algorithms.
result Achieves near-optimal sample complexity for finding approximate Nash equilibria.
New assumptions and algorithm solve offline two-player zero-sum Markov games.
problem Solving offline two-player zero-sum Markov games under insufficient assumptions.
method Proposed unilateral concentration assumption and pessimism-type algorithm.
result Algorithm efficiently learns Nash equilibrium under unilateral concentration.
Study best-response learning dynamics in zero-sum polymatrix games under full and minimal information settings.
problem Learning dynamics in zero-sum polymatrix games under different information settings.
method Two-timescale learning dynamics combining smoothed best-response updates and TD-learning for estimating local payoff functions.
result Polynomial-time finite-sample guarantees for convergence to an ε-Nash equilibrium in the minimal information case.
Expands MFGs to handle real-world asymmetric multi-agent games efficiently.
problem Applying mean-field games to real-world, heterogeneous multi-agent systems.
method Develops a method to symmetrize and extend finite-player games to infinite-player MFGs, proving approximation bounds and convergence guarantees.
result TD learning converges to approximate Nash equilibria in finite-sample settings, enabling symmetrized learning without explicit MFG models.
New algorithm finds near-optimal policies efficiently in zero-sum games.
problem Lack of provable efficiency guarantees for policy optimization in zero-sum games.
method Policy optimization algorithm with function approximation.
result Proves efficient convergence to near-optimal policies with polynomial samples and iterations.
In the standard setting of approachability there are two players and a target set. The players play repeatedly a known vector-valued game where the first player wants to have the average vector-valued payoff converge to the target set which the other player tries to exclude it from this set. We revisit this setting in …
We study stochastic multi-armed bandits with many players. The players do not know the number of players, cannot communicate with each other and if multiple players select a common arm they collide and none of them receive any reward. We consider the static scenario, where the number of players remains fixed, and the d…
New games model strategic interactions in incomplete information settings.
problem Modeling strategic interactions in incomplete information settings.
method Introduced new games that map input to private player types, aggregate strategies, and converge to near-Nash equilibria.
result Games can recover meaningful strategic interactions from real data.
A new algorithm improves sample complexity for thresholding in Monte Carlo Tree Search.
problem Determining if the root node value of a tree is at least a given threshold.
method Developed a δ-correct sequential sampling algorithm based on the Track-and-Stop strategy.
result Ratio-based modification of D-Tracking strategy reduces sample complexity and computational cost.
We consider a symmetric multi-players zero-sum game with two strategic variables. There are n n n players, n ≥ 3 n\geq 3 n ≥ 3 . Each player is denoted by i i i . Two strategic variables are t i t_i t i and s i s_i s i , i ∈ { 1 , … , n } i\in \{1, \dots, n\} i ∈ { 1 , … , n } . They are related by invertible functions. Using the minimax theorem by \cite{sion} we will show that Nas…
A policy for near-optimal multi-player bandits with non-zero collision rewards.
problem Decentralized multi-player bandits with heterogeneous rewards and collisions.
method A policy achieving near-optimal regret in a non-communicative setting.
result Near order-optimal expected regret of O ( log 1 + δ T ) O(\log^{1 + δ} T) O ( log 1 + δ T ) for 0 < δ < 1 0 < δ< 1 0 < δ < 1 . A multi-player bandit system resists adversarial attacks with near-optimal regret.
problem Adversaries attempt to manipulate rewards in a multi-player multi-armed bandit game.
method Players communicate a single bit to resist attacks, achieving near-optimal regret.
result Achieves near-optimal regret of O ( log 1 + δ T + W ) O(\log^{1+δ}T + W) O ( log 1 + δ T + W ) , where W W W is the total time of adversarial attacks. Paper optimizes reinforcement learning in self-play games with reduced steps.
problem Optimizing reinforcement learning algorithms for self-play in two-player zero-sum games.
method Proposes optimistic Nash Q-learning and Nash V-learning algorithms with improved sample complexity.
result Achieves sample complexity of O ( S A B ) O(SAB) O ( S A B ) for Nash Q-learning and O ( S ( A + B ) ) O(S(A+B)) O ( S ( A + B )) for Nash V-learning, closing the gap with lower bounds. Predict player lifetime value based on engagement metrics.
problem User profiling in video games.
method Survival curves, deep learning, long short-term memory.
result Deep learning predicts player lifetime value.
We consider two-player non-zero-sum stopping games in discrete time. Unlike Dynkin games, in our games the payoff of each player is revealed after both players stop. Moreover, each player can adjust her own stopping strategy according to the other player's action. In the first part of the paper, we consider the game wh…
In this paper we study the problem of exact recovery of the pure-strategy Nash equilibria (PSNE) set of a graphical game from noisy observations of joint actions of the players alone. We consider sparse linear influence games --- a parametric class of graphical games with linear payoffs, and represented by directed gra…
New algorithm reduces regret in multi-player bandits with collision information.
problem Optimizing decisions in multi-player bandits with collision penalties.
method Developed an algorithm with optimal T \sqrt{T} T regret under collision announcements, and sublinear regret without collision info. result First T \sqrt{T} T -type regret guarantee for non-stochastic multi-player multi-armed bandits with collision information. New algorithms for n-player games using a player-centered approach.
problem Learning near-perfect strategies in multiplayer games.
method Player-centered reinforcement learning with Final Adaptation RL.
result FARL is crucial for achieving near-perfect strategies in various games.
New algorithm for multi-player bandits with selfish players, achieving logarithmic regret.
problem Challenges of robustness to selfish players in multi-player bandits.
method First algorithm robust to selfish players achieving logarithmic regret, with or without collision observation.
result Achieved logarithmic regret for robust algorithms to selfish players in multi-player bandits.
We develop a machine learning approach to represent and analyze the underlying spatial structure that governs shot selection among professional basketball players in the NBA. Typically, NBA players are discussed and compared in an heuristic, imprecise manner that relies on unmeasured intuitions about player behavior. T…
Paper uses Apprenticeship Learning to model player behavior in interactive narratives.
problem Understanding and simulating player behavior in interactive narratives.
method Receding Horizon IRL (RHIRL) to learn reward functions and policies.
result RHIRL can learn action sequences and generate behavior similar to specific players.
New algorithm for multi-player bandits in changing environments.
problem Sequential action selection with collisions and adversarial losses.
method First provable Multi-player Bandit algorithm for changing environments.
result Resolves open problem in adversarial multi-player settings.
New algorithm tackles multi-player bandit problems with limited access to arms.
problem Limited access to dynamic local subsets of arms in multi-player multi-armed bandit problems.
method Adopted Upper Confidence Bound (UCB) for exploration-exploitation and distributed optimization for collisions.
result Proposes a decentralized algorithm with near-optimal regret guarantee.
Paper presents Transfer Portal model for accurate player performance predictions.
problem Predicting future player performance after a transfer.
method Personalized neural network and Bayesian updating framework.
result Model generates accurate predictions for player performance at new clubs.
A new algorithm RESYNC for defenders against malicious attackers in multi-player bandits.
problem Malicious players colliding with cooperative players to prevent rewards.
method Decentralized and robust algorithm RESYNC for defenders.
result RESYNC algorithm is order-optimal, performing gracefully as the number of collisions increases.
Study predicts soccer player market values using machine learning and SHAP for interpretability.
problem Predicting accurate market values for professional soccer players.
method Ensemble machine learning models, SHAP for interpretability, Boruta for feature selection.
result GBDT model achieved high predictive accuracy (R-squared 0.901, RMSE 3,221,632.175).
Algorithm optimizes multi-player learning with noisy rewards without direct communication.
problem Cooperative multi-player learning with noisy rewards and no communication.
method Upper and lower confidence bounds algorithm for optimal action selection.
result Achieves logarithmic O ( log T Δ a ) O(\frac{\log T}{Δ_{\bm{a}}}) O ( Δ a l o g T ) and O ( T log T ) O(\sqrt{T\log T}) O ( T log T ) regret. New strategy achieves optimal regret without communication or collisions in multi-player bandit.
problem Cooperative multi-player stochastic multi-armed bandit with shared randomness.
method Combination of combinatorial approach to generalize geometric intuition.
result Achieves near-optimal regret i l d e O ( T ) ilde{O}(\sqrt{T}) i l d e O ( T ) for any number of players and arms without collisions. Paper develops efficient algorithms for zero-sum Markov games with general function classes.
problem Challenging settings in zero-sum Markov games with parameterized value functions or models.
method Developed new model-free and model-based algorithms for decoupled and coordinated settings.
result Improved sample complexity and regret bounds for various settings.
New algorithm for multi-player bandits in decentralized, asynchronous systems.
problem Challenges in decentralized, asynchronous multi-player bandits, including coordination and player detection.
method Adaptive exploration-exploitation algorithm that reduces collisions and detects player presence.
result Achieves regret of O ( T log T + log T / Δ 2 ) \mathcal{O}(\sqrt{T \log T} + {\log T}/{Δ^2}) O ( T log T + log T / Δ 2 ) . First sample-efficient algorithm for learning EFCE in bandit feedback settings.
problem Learning EFCE in bandit feedback settings for IIEFGs.
method Proposed K K K -EFCE and uncoupled no-regret algorithm with wide-range regret minimization. result First sample-efficient algorithm for learning EFCE from bandit feedback.