Reinforcement Learning in Robotics

Inductive biases of neural specialization in spatial navigation

This study shows that specialized neural modules improve learning and generalization in navigation tasks, resembling primate behavior and using a Kalman filter-like state representation.

Inverse Reinforcement Learning to Study Motivation in Mouse Behavioral Paradigms

This paper applies inverse reinforcement learning to directly infer the underlying motivations driving mouse behavior in behavioral paradigms.

Synthesising integrated robot behaviour through reinforcement learning for homeostasis

This paper introduces an algorithm inspired by neuroscience to enable robots to exhibit integrated survival behaviors through reinforcement learning focused on homeostasis.

Neuron-level prediction and noise can implement flexible reward-seeking behavior

This paper demonstrates that neural networks can exhibit reward-seeking behavior through local predictive updates and internal noise, enabling autonomous and adaptable interaction with environments.

A meta reinforcement learning account of behavioral adaptation to volatility in recurrent neural networks

The paper shows that recurrent neural networks trained with medium or meta-volatility regimes adapt better to environmental volatility than those trained in low volatility.

Humans adaptively select different computational strategies in different learning environments.

The paper shows humans adaptively choose between flat and hierarchical learning strategies based on environmental complexity.

Generative Adversarial Implicit Successor Representation

This paper introduces a method to implicitly encode Successor Representations using a GAN, enabling action decisions without explicit probability calculations.

Goal Reduction with Loop-Removal Accelerates RL and Models Human Brain Activity in Goal-Directed Learning

This paper introduces a goal reduction mechanism that accelerates reinforcement learning and models human brain activity in goal-directed learning by deriving subgoals through loop-removal.

Emergence of belief-like representations through reinforcement learning

This paper shows that recurrent neural networks can estimate value directly from observations without explicitly forming beliefs, generating reward prediction errors similar to experimental observations.

Having "multiple selves" helps learning agents explore and adapt in complex changing worlds

This paper shows that modular agents, composed of subagents for distinct needs, outperform monolithic agents in maintaining homeostasis in changing environments.

CoBeL-RL: A neuroscience-oriented simulation framework for complex behavior and learning

CoBeL-RL is a simulation framework for modeling complex behavior and learning in neuroscience, offering virtual environments and RL algorithms for efficient simulation and analysis.

Combining Backpropagation with Equilibrium Propagation to improve an Actor-Critic RL framework

This paper combines equilibrium propagation with backpropagation in an actor-critic reinforcement learning framework, showing improved consistency in high-reward behavior on control tasks.

Reward Bases: Instantaneous reward revaluation with temporal difference learning

The paper introduces an extension to temporal difference learning that enables instantaneous adaptation to changing reward functions without retraining.

Validating the Representational Space of DeepReinforcement Learning Models of Behavior withNeural Data

This paper validates the neural representations of Deep RL agents against animal neural data, showing similar geometry in latent spaces and the role of temporal information.

Novelty is not Surprise: Human exploratory and adaptive behavior in sequential decision-making

This study introduces a deep sequential decision-making paradigm to explain human behavior by incorporating novelty and surprise, showing that model-free actions dominate decisions despite the availability of a world-model.

Temporal and state abstractions for efficient learning, transfer and composition in humans

This study investigates how humans use hierarchical temporal and state abstractions to transfer knowledge across tasks, showing transfer effects not explained by simpler models.

Optimal Reinforcement Learning with Asymmetric Updating in Volatile Environments: a Simulation Study

This study compares two reinforcement learning models in volatile environments, finding that separate reward and punishment learning rates improve performance, with higher reward learning rates being more beneficial.

What is the state space of the world for real animals?

The paper examines assumptions about state spaces in reinforcement learning models applied to animal behavior and proposes a continuous time Markov renewal process model to address limitations.

Multi-Task Reinforcement Learning in Humans

This study investigates human multi-task reinforcement learning by comparing participant behavior to algorithms in a two-step decision-making task with changing rewards.

Habituation and Goal-directed arbitration mechanisms and failures under partial observability

The paper models habituation and goal-directed behavior arbitration under partial observability using reinforcement learning to identify new mechanisms and failures in decision-making processes.

Computational evidence for hierarchically-structured reinforcement learning in humans

This study shows human behavior in complex environments can be explained by hierarchical reinforcement learning, outperforming flat RL and hierarchical Bayesian models in capturing observed patterns.

Computational neural mechanisms of goal-directed planning and problem solving

This paper introduces a localist neural network model that autonomously learns environment structures and achieves arbitrary goals without re-learning reward values, demonstrating its ability to solve complex problems like the Tower of Hanoi.

Biological Reinforcement Learning via Predictive Spacetime Encoding

This paper introduces a novel reinforcement learning model inspired by hippocampal and prefrontal cortex mechanisms, showing improved performance in uncertain environments.

Computational characteristics of the striatal dopamine system described by reinforcement learning with fast generalization

This paper introduces a reinforcement learning model inspired by the striatal dopamine system, showing it enables fast generalization and safe exploration but can also lead to maladaptive behaviors.

Imitation as a model-free process in human reinforcement learning

This study tests three hypotheses about how imitation influences human reinforcement learning and finds that value shaping is the most supported model.

Generalizing to generalize: humans flexibly switch between compositional and conjunctive structures during reinforcement learning

The paper investigates how humans adaptively generalize task structures during reinforcement learning, finding they switch between compositional and conjunctive generalization based on task statistics.

A Comparison of Non-human Primate and Deep Reinforcement Learning Agent Performance in a Virtual Pursuit-Avoidance Task

This study compares non-human primate and deep reinforcement learning agent performance in a virtual pursuit-avoidance task, finding agents primarily learn first-order physics while the primate exhibits second-order physics.

Computational mechanisms of curiosity and goal-directed exploration

The paper presents two computational mechanisms for goal-directed exploration based on different types of information-gain, showing how they induce distinct patterns of Bayes-optimal behavior.

Learning and Forgetting Using Reinforced Bayesian Change Detection

This paper introduces a Bayesian learning model with adaptive forgetting to improve reinforcement learning by dynamically adjusting to environmental changes and balancing exploration-exploitation.

Selective Maintenance of Value Information Helps Resolve the Exploration/Exploitation Dilemma

This paper shows that selectively maintaining high-value information helps resolve the exploration/exploitation dilemma in decision-making, performing as well as uncertainty-driven exploration but with lower cognitive load.