Financial exchanges across the world use limit order books (LOBs) to pro...
In many real world settings binary classification decisions are made bas...
Communication is a powerful tool for coordination in multi-agent RL. But...
Training multiple agents to coordinate is an important problem with
appl...
Meta-learning, the notion of learning to learn, enables learning systems...
Structured state space sequence (S4) models have recently achieved
state...
Open-ended learning methods that automatically generate a curriculum of
...
As machine learning agents act more autonomously in the world, they will...
Adversarial attacks in reinforcement learning (RL) often assume
highly-p...
Steganography is the practice of encoding secret information into innocu...
Successful coordination in Dec-POMDPs requires agents to adopt robust
st...
Tremendous progress has been made in reinforcement learning (RL) over th...
We consider the problem of making AI agents that collaborate well with h...
Meta-gradients provide a general approach for optimizing the meta-parame...
Adaptive curricula in reinforcement learning (RL) have proven effective ...
Self-play is a common paradigm for constructing solutions in Markov game...
In general-sum games, the interaction of self-interested learning agents...
Learning in general-sum games can be unstable and often leads to sociall...
It remains a significant challenge to train generally capable agents wit...
An unaddressed challenge in human-AI coordination is to enable AI agents...
General policy improvement (GPI) and trust-region learning (TRL) are the...
Ridge Rider (RR) is an algorithm for finding diverse solutions to
optimi...
We study a class of classification problems best exemplified by the
bank...
Deep reinforcement learning (RL) agents may successfully generalize to n...
Self-supervised pre-training of large-scale transformer models on text
c...
In many common-payoff games, achieving good performance requires players...
Reinforcement learning (RL) in partially observable, fully cooperative
m...
Search is an important tool for computing effective policies in single- ...
In many coordination problems, independently reasoning humans are able t...
Effective communication is an important skill for enabling information
e...
The standard problem setting in Dec-POMDPs is self-play, where the goal ...
Over the last decade, a single algorithm has changed many facets of our ...
Effective communication is an important skill for enabling information
e...
For neural models to garner widespread public trust and ensure fairness,...
In many real-world settings, a team of agents must coordinate its behavi...
We consider the problem of zero-shot coordination - constructing AI agen...
Recent superhuman results in games have largely been achieved in a varie...
Many recent works have discussed the propensity, or lack thereof, for
em...
Gradient-based methods for optimisation of objectives in stochastic sett...
To be successful in real-world tasks, Reinforcement Learning (RL) needs ...
Deep learning is built on the foundational guarantee that gradient desce...
How do we know if communication is emerging in a multi-agent system? The...
In the last few years, deep multi-agent reinforcement learning (RL) has
...
A growing number of learning methods are actually games which optimise
m...
We present Pommerman, a multi-agent environment based on the classic con...
In many real-world settings, a team of agents must coordinate their beha...
The cornerstone underpinning deep learning is the guarantee that gradien...
The score function estimator is widely used for estimating gradients of
...
We model the spread of news as a social learning game on a network. Agen...
Cooperative multi-agent systems can be naturally used to model many real...