Search Results for author: Murray Shanahan

Found 32 papers, 10 papers with code

Abstraction for Deep Reinforcement Learning

no code implementations10 Feb 2022 Murray Shanahan, Melanie Mitchell

We characterise the problem of abstraction in the context of deep reinforcement learning.

reinforcement-learning

Feature-Attending Recurrent Modules for Generalizing Object-Centric Behavior

no code implementations15 Dec 2021 Wilka Carvalho, Andrew Lampinen, Kyriacos Nikiforou, Felix Hill, Murray Shanahan

To generalize in object-centric tasks, a reinforcement learning (RL) agent needs to exploit the structure that objects induce.

In a Nutshell, the Human Asked for This: Latent Goals for Following Temporal Specifications

2 code implementations ICLR 2022 Borja G. León, Murray Shanahan, Francesco Belardinelli

We address the problem of building agents whose goal is to learn to execute out-of distribution (OOD) multi-task instructions expressed in temporal logic (TL) by using deep reinforcement learning (DRL).

reinforcement-learning

Ensembles and Encoders for Task-Free Continual Learning

no code implementations29 Sep 2021 Murray Shanahan, Christos Kaplanis, Jovana Mitrović

We present an architecture that is effective for continual learning in an especially demanding setting, where task boundaries do not exist or are unknown, and where classes have to be learned online (with each presented only once).

Continual Learning Self-Supervised Learning

Task-driven Discovery of Perceptual Schemas for Generalization in Reinforcement Learning

no code implementations29 Sep 2021 Wilka Torrico Carvalho, Andrew Kyle Lampinen, Kyriacos Nikiforou, Felix Hill, Murray Shanahan

Taking inspiration from cognitive science, we term representations for reoccurring segments of an agent's experience, "perceptual schemas".

reinforcement-learning

Integrated information as a common signature of dynamical and information-processing complexity

no code implementations18 Jun 2021 Pedro A. M. Mediano, Fernando E. Rosas, Juan Carlos Farah, Murray Shanahan, Daniel Bor, Adam B. Barrett

The apparent dichotomy between information-processing and dynamical approaches to complexity science forces researchers to choose between two diverging sets of tools and explanations, creating conflict and often hindering scientific progress.

Learning to Represent State with Perceptual Schemata

no code implementations ICML Workshop URL 2021 Wilka Torrico Carvalho, Murray Shanahan

We present empirical results that Perceptual Schemata enables a state representation that can maintain multiple objects observed in sequence with independent dynamics while an LSTM cannot.

Encoders and Ensembles for Task-Free Continual Learning

no code implementations27 May 2021 Murray Shanahan, Christos Kaplanis, Jovana Mitrović

We present an architecture that is effective for continual learning in an especially demanding setting, where task boundaries do not exist or are unknown, and where classes have to be learned online (with each example presented only once).

Continual Learning Image Classification +1

Unsupervised Object-Based Transition Models for 3D Partially Observable Environments

no code implementations NeurIPS 2021 Antonia Creswell, Rishabh Kabra, Chris Burgess, Murray Shanahan

We present a slot-wise, object-based transition model that decomposes a scene into objects, aligns them (with respect to a slot-wise object memory) to maintain a consistent order across time, and predicts how those objects evolve over successive frames.

Learning a Non-Redundant Collection of Classifiers

no code implementations1 Jan 2021 Daniel Pace, Alessandra Russo, Murray Shanahan

Inspired by Quality-Diversity algorithms, in this work we train a collection of classifiers to learn distinct solutions to a classification problem, with the goal of learning to exploit a variety of predictive signals present in the training data.

Learning to Combine Top-Down and Bottom-Up Signals in Recurrent Neural Networks with Attention over Modules

1 code implementation ICML 2020 Sarthak Mittal, Alex Lamb, Anirudh Goyal, Vikram Voleti, Murray Shanahan, Guillaume Lajoie, Michael Mozer, Yoshua Bengio

To effectively utilize the wealth of potential top-down information available, and to prevent the cacophony of intermixed signals in a bidirectional architecture, mechanisms are needed to restrict information flow.

Language Modelling Sequential Image Classification +1

Learning Diverse Representations for Fast Adaptation to Distribution Shift

no code implementations12 Jun 2020 Daniel Pace, Alessandra Russo, Murray Shanahan

assumption is a useful idealization that underpins many successful approaches to supervised machine learning.

Systematic Generalisation through Task Temporal Logic and Deep Reinforcement Learning

no code implementations12 Jun 2020 Borja G. León, Murray Shanahan, Francesco Belardinelli

This work introduces a neuro-symbolic agent that combines deep reinforcement learning (DRL) with temporal logic (TL) to achieve systematic zero-shot, i. e., never-seen-before, generalisation of formally specified instructions.

reinforcement-learning

Continual Reinforcement Learning with Multi-Timescale Replay

1 code implementation16 Apr 2020 Christos Kaplanis, Claudia Clopath, Murray Shanahan

In this paper, we propose a multi-timescale replay (MTR) buffer for improving continual learning in RL agents faced with environments that are changing continuously over time at timescales that are unknown to the agent.

Continual Learning Continuous Control +1

AlignNet: Self-supervised Alignment Module

no code implementations25 Sep 2019 Antonia Creswell, Luis Piloto, David Barrett, Kyriacos Nikiforou, David Raposo, Marta Garnelo, Peter Battaglia, Murray Shanahan

The natural world consists of objects that we perceive as persistent in space and time, even though these objects appear, disappear and reappear in our field of view as we move.

Question Answering

The Animal-AI Environment: Training and Testing Animal-Like Artificial Cognition

4 code implementations12 Sep 2019 Benjamin Beyret, José Hernández-Orallo, Lucy Cheke, Marta Halina, Murray Shanahan, Matthew Crosby

Recent advances in artificial intelligence have been strongly driven by the use of game environments for training and evaluating agents.

An Explicitly Relational Neural Network Architecture

no code implementations ICML 2020 Murray Shanahan, Kyriacos Nikiforou, Antonia Creswell, Christos Kaplanis, David Barrett, Marta Garnelo

With a view to bridging the gap between deep learning and symbolic AI, we present a novel end-to-end neural network architecture that learns to form propositional representations with an explicitly relational structure from raw pixel data.

Relational Reasoning

Deep reinforcement learning with relational inductive biases

no code implementations ICLR 2019 Vinicius Zambaldi, David Raposo, Adam Santoro, Victor Bapst, Yujia Li, Igor Babuschkin, Karl Tuyls, David Reichert, Timothy Lillicrap, Edward Lockhart, Murray Shanahan, Victoria Langston, Razvan Pascanu, Matthew Botvinick, Oriol Vinyals, Peter Battaglia

We introduce an approach for augmenting model-free deep reinforcement learning agents with a mechanism for relational reasoning over structured representations, which improves performance, learning efficiency, generalization, and interpretability.

reinforcement-learning Relational Reasoning +2

Consistent Jumpy Predictions for Videos and Scenes

no code implementations ICLR 2019 Ananya Kumar, S. M. Ali Eslami, Danilo Rezende, Marta Garnelo, Fabio Viola, Edward Lockhart, Murray Shanahan

These models typically generate future frames in an autoregressive fashion, which is slow and requires the input and output frames to be consecutive.

3D Scene Reconstruction Video Prediction

Policy Consolidation for Continual Reinforcement Learning

1 code implementation1 Feb 2019 Christos Kaplanis, Murray Shanahan, Claudia Clopath

We propose a method for tackling catastrophic forgetting in deep reinforcement learning that is \textit{agnostic} to the timescale of changes in the distribution of experiences, does not require knowledge of task boundaries, and can adapt in \textit{continuously} changing environments.

Continual Learning Continuous Control +1

Explicit Information Placement on Latent Variables using Auxiliary Generative Modelling Task

no code implementations27 Sep 2018 Nat Dilokthanakul, Nick Pawlowski, Murray Shanahan

We demonstrate the use of the method in a task of disentangling global structure from local features in images.

Consistent Generative Query Networks

1 code implementation ICLR 2019 Ananya Kumar, S. M. Ali Eslami, Danilo J. Rezende, Marta Garnelo, Fabio Viola, Edward Lockhart, Murray Shanahan

These models typically generate future frames in an autoregressive fashion, which is slow and requires the input and output frames to be consecutive.

3D Scene Reconstruction Video Prediction

Relational Deep Reinforcement Learning

7 code implementations5 Jun 2018 Vinicius Zambaldi, David Raposo, Adam Santoro, Victor Bapst, Yujia Li, Igor Babuschkin, Karl Tuyls, David Reichert, Timothy Lillicrap, Edward Lockhart, Murray Shanahan, Victoria Langston, Razvan Pascanu, Matthew Botvinick, Oriol Vinyals, Peter Battaglia

We introduce an approach for deep reinforcement learning (RL) that improves upon the efficiency, generalization capacity, and interpretability of conventional approaches through structured perception and relational reasoning.

reinforcement-learning Relational Reasoning +2

Continual Reinforcement Learning with Complex Synapses

no code implementations ICML 2018 Christos Kaplanis, Murray Shanahan, Claudia Clopath

Unlike humans, who are capable of continual learning over their lifetimes, artificial neural networks have long been known to suffer from a phenomenon known as catastrophic forgetting, whereby new learning can lead to abrupt erasure of previously acquired knowledge.

Continual Learning reinforcement-learning

SCAN: Learning Hierarchical Compositional Visual Concepts

no code implementations ICLR 2018 Irina Higgins, Nicolas Sonnerat, Loic Matthey, Arka Pal, Christopher P. Burgess, Matko Bosnjak, Murray Shanahan, Matthew Botvinick, Demis Hassabis, Alexander Lerchner

SCAN learns concepts through fast symbol association, grounding them in disentangled visual primitives that are discovered in an unsupervised manner.

Feature Control as Intrinsic Motivation for Hierarchical Reinforcement Learning

1 code implementation18 May 2017 Nat Dilokthanakul, Christos Kaplanis, Nick Pawlowski, Murray Shanahan

We highlight the advantage of our approach in one of the hardest games -- Montezuma's revenge -- for which the ability to handle sparse rewards is key.

Hierarchical Reinforcement Learning Montezuma's Revenge +1

Deep Unsupervised Clustering with Gaussian Mixture Variational Autoencoders

3 code implementations8 Nov 2016 Nat Dilokthanakul, Pedro A. M. Mediano, Marta Garnelo, Matthew C. H. Lee, Hugh Salimbeni, Kai Arulkumaran, Murray Shanahan

We study a variant of the variational autoencoder model (VAE) with a Gaussian mixture as a prior distribution, with the goal of performing unsupervised clustering through deep generative models.

Towards Deep Symbolic Reinforcement Learning

no code implementations18 Sep 2016 Marta Garnelo, Kai Arulkumaran, Murray Shanahan

Deep reinforcement learning (DRL) brings the power of deep neural networks to bear on the generic task of trial-and-error learning, and its effectiveness has been convincingly demonstrated on tasks such as Atari video games and the game of Go.

Game of Go reinforcement-learning +1

Classifying Options for Deep Reinforcement Learning

no code implementations27 Apr 2016 Kai Arulkumaran, Nat Dilokthanakul, Murray Shanahan, Anil Anthony Bharath

In this paper we combine one method for hierarchical reinforcement learning - the options framework - with deep Q-networks (DQNs) through the use of different "option heads" on the policy network, and a supervisory network for choosing between the different options.

Hierarchical Reinforcement Learning reinforcement-learning

Ascribing Consciousness to Artificial Intelligence

no code implementations22 Apr 2015 Murray Shanahan

This paper critically assesses the anti-functionalist stance on consciousness adopted by certain advocates of integrated information theory (IIT), a corollary of which is that human-level artificial intelligence implemented on conventional computing hardware is necessarily not conscious.

Cannot find the paper you are looking for? You can Submit a new open access paper.