A new reinforcement learning method uses mutual information to encourage agents to control their environment.
problem Learning from internal drives instead of external rewards.
method Formulate an intrinsic objective as mutual information between goal states and controllable states, derive a surrogate objective for efficient optimization.
result Demonstrated the efficacy of the approach in robotic tasks.
New method learns diverse solutions in reinforcement learning without gradient bias.
problem Lack of diverse solutions in reinforcement learning tasks.
method Maximizes state-action-based mutual information directly, using variational lower bound.
result Successfully learns an infinite set of diverse solutions.
Exploration is a difficult challenge in reinforcement learning and is of prime importance in sparse reward environments. However, many of the state of the art deep reinforcement learning algorithms, that rely on epsilon-greedy, fail on these environments. In such cases, empowerment can serve as an intrinsic reward sign…
MIME uses mutual information minimization for better exploration in environments with abrupt transitions.
problem Agents struggle at abrupt environmental transitions.
method MIME learns a latent representation without predicting future states.
result MIME outperforms surprisal-driven agents at transition boundaries.
Our work proves CSF can recover ground-truth features in RL, improving understanding of feature learning.
problem Understanding the role of representation and mutual information in reinforcement learning.
method Investigates Contrastive Successor Features (CSF) method for identifiable representation learning in reinforcement learning.
result Proves CSF can recover ground-truth features up to a linear transformation.
Mutual information maximization has emerged as a powerful learning objective for unsupervised representation learning obtaining state-of-the-art performance in applications such as object recognition, speech recognition, and reinforcement learning. However, such approaches are fundamentally limited since a tight lower …
A new method for dynamic feature selection outperforms existing approaches.
problem Sequentially selecting features based on current information in machine learning.
method Greedy selection of features based on conditional mutual information, combined with a learning approach for optimization.
result The method outperforms existing feature selection methods in experiments.
MIRO learns robust latent spaces by maximizing mutual information with future information.
problem Robust perception in complex, unstructured environments with low sample complexity.
method MIRO maximizes mutual information in a latent space for model-based reinforcement learning.
result MIRO outperforms reconstruction objectives in cluttered scenes.
Learning to cooperate is crucially important in multi-agent environments. The key is to understand the mutual interplay between agents. However, multi-agent environments are highly dynamic, where agents keep moving and their neighbors change quickly. This makes it hard to learn abstract representations of mutual interp…
New method calculates empowerment from visual data, solving complex reinforcement learning problems.
problem Calculating empowerment in unknown dynamics from visual observation is challenging.
method Developed a novel approach using stochastic dynamic models in latent space and the Water-Filling algorithm.
result Efficiently computed empowerment in unknown dynamics from visual observation only.
The mutual information is a core statistical quantity that has applications in all areas of machine learning, whether this is in training of density models over multiple data modalities, in maximising the efficiency of noisy transmission channels, or when learning behaviour policies for exploration by artificial agents…
When estimating the relevancy between a query and a document, ranking models largely neglect the mutual information among documents. A common wisdom is that if two documents are similar in terms of the same query, they are more likely to have similar relevance score. To mitigate this problem, in this paper, we propose …
Learning to cooperate with friends and compete with foes is a key component of multi-agent reinforcement learning. Typically to do so, one requires access to either a model of or interaction with the other agent(s). Here we show how to learn effective strategies for cooperation and competition in an asymmetric informat…
Recurrent networks learn beliefs from history in partially observable environments.
problem Learning optimal policies in partially observable environments.
method Trained recurrent neural networks to approximate value functions, measuring mutual information between hidden states and beliefs.
result Recurrent networks' hidden states correlate with beliefs of relevant state variables, improving expected return.
Study privacy-utility trade-off in IoT time-series data sharing with RL.
problem Privacy concerns in IoT time-series data sharing with temporal correlations.
method Reformulated as MDP, solved with asynchronous actor-critic deep RL.
result Validated solution on synthetic and real GPS datasets.
A deep reinforcement learning approach for slate re-ranking in e-commerce.
problem Improving user satisfaction in e-commerce by optimizing the ranking of items in a slate.
method Generator and Critic approach, using reinforcement learning and a Full Slate Critic model.
result The Generator and Critic approach significantly outperforms existing methods in slate evaluation and efficiency.
Study reveals mutual reinforcement between adversarial inputs and poisoned models.
problem Understanding and mitigating vulnerabilities of deep learning models.
method Developed a new attack model to jointly optimize adversarial inputs and poisoned models.
result Mutual reinforcement effects between adversarial inputs and poisoned models significantly amplify each other's effectiveness.
Reinforcement learning for embodied agents is a challenging problem. The accumulated reward to be optimized is often a very rugged function, and gradient methods are impaired by many local optimizers. We demonstrate, in an experimental setting, that incorporating an intrinsic reward can smoothen the optimization landsc…
Proposes PIC and POIC for measuring task difficulty in RL.
problem Lack of metrics to measure task difficulty in RL.
method Introduces policy information capacity (PIC) and policy-optimal information capacity (POIC) as metrics based on mutual information.
result Empirically shows PIC and POIC correlate with task solvability better than alternatives.
CVRL tackles complex visual observations in reinforcement learning.
problem Complex visual observations in natural environments.
method Contrastive Variational Reinforcement Learning (CVRL) learns a contrastive variational model by maximizing mutual information between latent states and observations.
result CVRL achieves comparable performance with state-of-the-art model-based DRL methods and significantly outperforms them on tasks with complex observations.
New framework for RL transfer learning with state-action mismatch.
problem High sample complexity in RL from scratch.
method Embeddings to transfer knowledge between MDPs with different state- and action-spaces.
result Successful transfer learning in scenarios with state- and action-space mismatches.
CARMS improves gradient estimation for categorical variables.
problem Accurately backpropagating gradients through categorical variables.
method CARMS combines REINFORCE with antithetic sampling to create unbiased gradient estimators.
result CARMS outperforms competing methods on various tasks.
In industrial environments, an increasing amount of wireless devices are used, which utilize license-free bands. As a consequence of these mutual interferences of wireless systems might decrease the state of coexistence. Therefore, a central coexistence management system is needed, which allocates conflict-free resourc…
ARMS improves gradient estimation for binary variables using antithetic samples.
problem Estimating gradients for binary variables in discrete latent variable models.
method ARMS uses antithetic samples generated by a copula to estimate gradients more efficiently and unbiasedly.
result ARMS outperforms competing methods in training generative models and optimizing variational bounds.
Meta-learning algorithms use past experience to learn to quickly solve new tasks. In the context of reinforcement learning, meta-learning algorithms acquire reinforcement learning procedures to solve new problems more efficiently by utilizing experience from prior tasks. The performance of meta-learning algorithms depe…
Cumulative entropy regularization introduces a regulatory signal to the reinforcement learning (RL) problem that encourages policies with high-entropy actions, which is equivalent to enforcing small deviations from a uniform reference marginal policy. This has been shown to improve exploration and robustness, and it ta…
New insights link RLHF and contrastive learning for better model alignment.
problem Aligning large language models with human values.
method Interpreting RLHF and DPO as contrastive learning methods based on mutual information.
result Proposed Mutual Information Optimization (MIO) improves model performance.
Enhances RL agents with predictive internal representations.
problem Improving model-free reinforcement learning agents.
method Introduces Deep InfoMax (DIM) objective to train predictive internal representations.
result Successfully learned predictive representations in synthetic settings.
This paper introduces a new framework for learning nearly decomposable Q-functions via communication minimization.
problem Challenges in multi-agent reinforcement learning, especially scalability and non-stationarity.
method Learning nearly decomposable Q-functions (NDQ) via communication minimization, introducing two information-theoretic regularizers.
result Significantly outperforms baseline methods on the StarCraft unit micromanagement benchmark.
New approach uses SPG for semantic communication without a known channel model.
problem Designing efficient semantic communication systems without a known channel model.
method Applying Stochastic Policy Gradient (SPG) for reinforcement learning.
result Achieves comparable performance to model-aware approaches with a decreased convergence rate.
We consider a framework involving behavioral economics and machine learning. Rationally inattentive Bayesian agents make decisions based on their posterior distribution, utility function and information acquisition cost Renyi divergence which generalizes Shannon mutual information). By observing these decisions, how ca…
CausalCOMRL improves RL task representations by integrating causal relationships, enhancing generalizability.
problem Spurious correlations in context-based offline meta-reinforcement learning.
method CausalCOMRL integrates causal representation learning to uncover and incorporate causal relationships among task components.
result CausalCOMRL achieves better performance on meta-reinforcement learning benchmarks.
We propose a new method to study the internal memory used by reinforcement learning policies. We estimate the amount of relevant past information by estimating mutual information between behavior histories and the current action of an agent. We perform this estimation in the passive setting, that is, we do not interven…
A novel method for visual question answering using scene graphs and reinforcement learning.
problem Answering free-form questions about images with deep linguistic and visual understanding.
method Context-driven, sequential reasoning based on scene graphs and reinforcement learning.
result Our method almost reaches human performance on the GQA dataset.
Real-world tasks are often highly structured. Hierarchical reinforcement learning (HRL) has attracted research interest as an approach for leveraging the hierarchical structure of a given task in reinforcement learning (RL). However, identifying the hierarchical policy structure that enhances the performance of RL is n…
Action-bisimulation learns long-horizon controllability for reinforcement learning.
problem Learning relevant state features in high-dimensional observations for robust reinforcement learning.
method Action-bisimulation encoding, inspired by bisimulation invariance, extends single-step controllability to multi-step.
result Action-bisimulation pretraining improves sample efficiency in various environments.
In reinforcement learning (RL), temporal abstraction still remains as an important and unsolved problem. The options framework provided clues to temporal abstraction in the RL, and the option-critic architecture elegantly solved the two problems of finding options and learning RL agents in an end-to-end manner. However…
New framework for modular reinforcement learning reduces sample complexity.
problem Achieving independent credit assignment in reinforcement learning.
method Defining modular credit assignment as minimizing algorithmic mutual information, introducing modularity criterion for causal analysis.
result Single-step temporal difference action-value methods meet the modularity criterion, improving sample efficiency.
Two new exploration methods for multi-agent systems improve team performance.
problem Exploration in transition-dependent multi-agent settings.
method EITI and EDTI, using mutual information and VoI to encourage coordinated exploration.
result Significant improvement in multi-agent performance through coordinated exploration.
Reinforcement Learning (RL) has achieved impressive performance in many complex environments due to the integration with Deep Neural Networks (DNNs). At the same time, Genetic Algorithms (GAs), often seen as a competing approach to RL, had limited success in scaling up to the DNNs required to solve challenging tasks. C…
New neural network approach using mutual information.
problem Training neural networks for imbalanced datasets.
method Converts neural network classifiers to mutual information evaluators.
result New form of softmax leads to better classification accuracy, especially for imbalanced datasets.
The study explains YouTube commenters' behavior using rational inattention models.
problem Understanding and predicting YouTube commenters' behavior.
method Deep embedded clustering for user grouping, Bayesian revealed preferences for rationality testing, and behavioral economics constraints for attention span modeling.
result Most YouTube user groups optimize a Bayesian utility with rationally inattentive constraints.
Paper describes profiles of multivariate normal distributions and novel estimators for mutual information.
problem Estimating mutual information for complex distributions.
method Analytical description of profiles, introduction of Bend and Mix Models, Monte Carlo estimation.
result Bend and Mix Models accurately estimate mutual information profiles and provide Bayesian estimates.
Paper benchmarks mutual info estimators on diverse distributions.
problem Evaluating mutual information estimators on complex, real-world distributions.
method Constructs a diverse family of known-ground truth distributions, proposes a benchmark platform.
result Highlights differences in classical and neural estimators' performance across various conditions.
Neural estimator improves mutual information estimation in high dimensions.
problem Estimating mutual information in high dimensions is challenging.
method Parametrizing conditional densities with normalizing flows and using block autoregressive structure.
result Improved mutual information estimation on benchmark tasks.
Improved bounds on learning algorithms' performance using conditional mutual information.
problem Bounding the generalization error of learning algorithms.
method Introducing conditional mutual information and disintegrated mutual information to tighten bounds.
result New bounds are tighter than previous ones, especially for noisy, iterative algorithms.
We propose three measures of mutual dependence between multiple random vectors. All the measures are zero if and only if the random vectors are mutually independent. The first measure generalizes distance covariance from pairwise dependence to mutual dependence, while the other two measures are sums of squared distance…
The paper argues that normalized mutual information is biased in clustering and community detection.
problem Bias in normalized mutual information for clustering and community detection.
method Introducing a modified version of mutual information to correct for information content and spurious dependence.
result The modified mutual information leads to different conclusions about which algorithms are best for community detection.