Learning Dynamics Model in Reinforcement Learning by Incorporating the Long Term Future
Nan Rosemary Ke
Amanpreet Singh
Ahmed Touati
Anirudh Goyal
Yoshua Bengio
Devi Parikh
Dhruv Batra
In model-based reinforcement learning, the agent interleaves between model learning and planning. These two components are inextricably inte… (see more)rtwined. If the model is not able to provide sensible long-term prediction, the executed planner would exploit model flaws, which can yield catastrophic failures. This paper focuses on building a model that reasons about the long-term future and demonstrates how to use this for efficient planning and exploration. To this end, we build a latent-variable autoregressive model by leveraging recent ideas in variational inference. We argue that forcing latent variables to carry future information through an auxiliary task substantially improves long-term predictions. Moreover, by planning in the latent space, the planner's solution is ensured to be within regions where the model is valid. An exploration strategy can be devised by searching for unlikely trajectories under the model. Our method achieves higher reward faster compared to baselines on a variety of tasks and environments in both the imitation learning and model-based reinforcement learning settings.
2019-03-05
ArXiv (preprint)
arxiv.org
arxiv.org
Reinforcement Learning in Stationary Mean-field Games
Jayakumar Subramanian
Aditya Mahajan
Multi-agent reinforcement learning has made significant progress in recent years, but it remains a hard problem. Hence, one often resorts to… (see more) developing learning algorithms for specific classes of multi-agent systems. In this paper we study reinforcement learning in a specific class of multi-agent systems systems called mean-field games. In particular, we consider learning in stationary mean-field games. We identify two different solution concepts---stationary mean-field equilibrium and stationary mean-field social-welfare optimal policy---for such games based on whether the agents are non-cooperative or cooperative, respectively. We then generalize these solution concepts to their local variants using bounded rationality based arguments. For these two local solution concepts, we present two reinforcement learning algorithms. We show that the algorithms converge to the right solution under mild technical conditions and demonstrate this using two numerical examples.
2019-03-01
Adaptive Agents and Multi-Agent Systems (published)
dblp.uni-trier.de
Stochastic Bit-Wise Iterative Decoding of Polar Codes
Kaining Han
Junchao Wang
Warren Gross
Jianhao Hu
Polar codes have received recent attention due to their potential to be applied in advanced wireless communication protocols such as the fif… (see more)th generation mobile communication system (5G). Among the existing decoding algorithms, Belief Propagation (BP) exhibits high-throughput, low-latency, and soft output with a high hardware cost. Stochastic computing, as a form of approximate computing, provides a potential low-cost implementation solution for the BP algorithm. However, existing stochastic BP decoders suffer from a relatively long decoding latency resulting in low hardware efficiency. In this paper, a novel bit-wise iterative stochastic decoding architecture for the BP algorithm is proposed to improve the throughput and hardware efficiency. By utilizing the frozen bits of polar codes and stochastic computing, multiple novel optimization methods are presented to further speed up convergence and increase the hardware efficiency.
2019-03-01
IEEE Transactions on Signal Processing (published)
doi.org
Stochastic Bit-Wise Iterative Decoding of Polar Codes
Kaining Han
Junchao Wang
Warren Gross
Jianhao Hu
Polar codes have received recent attention due to their potential to be applied in advanced wireless communication protocols such as the fif… (see more)th generation mobile communication system (5G). Among the existing decoding algorithms, Belief Propagation (BP) exhibits high-throughput, low-latency, and soft output with a high hardware cost. Stochastic computing, as a form of approximate computing, provides a potential low-cost implementation solution for the BP algorithm. However, existing stochastic BP decoders suffer from a relatively long decoding latency resulting in low hardware efficiency. In this paper, a novel bit-wise iterative stochastic decoding architecture for the BP algorithm is proposed to improve the throughput and hardware efficiency. By utilizing the frozen bits of polar codes and stochastic computing, multiple novel optimization methods are presented to further speed up convergence and increase the hardware efficiency.
2019-03-01
IEEE Transactions on Signal Processing (published)
doi.org
Prediction of Progression in Multiple Sclerosis Patients
Adrian Tousignant
Paul Lemaitre
Doina Precup
Douglas Arnold
Tal Arbel
We present the first automatic end-to-end deep learning framework for the prediction of future patient disability progression (one year from… (see more) baseline) based on multi-modal brain Magnetic Resonance Images (MRI) of patients with Multiple Sclerosis (MS). The model uses parallel convolutional pathways, an idea introduced by the popular Inception net and is trained and tested on two large proprietary, multi-scanner, multi-center, clinical trial datasets of patients with Relapsing-Remitting Multiple Sclerosis (RRMS). Experiments on 465 patients on the placebo arms of the trials indicate that the model can accurately predict future disease progression, measured by a sustained increase in the extended disability status scale (EDSS) score over time. Using only the multi-modal MRI provided at baseline, the model achieves an AUC of 0.66 +- 0.055. However, when supplemental lesion label masks are provided as inputs as well, the AUC increases to 0.701 +- 0.027. Furthermore, we demonstrate that uncertainty estimates based on Monte Carlo dropout sample variance correlate with errors made by the model. Clinicians provided with the predictions computed by the model can therefore use the associated uncertainty estimates to assess which scans require further examination.
2019-02-28
MIDL.io/2019/Conference (poster)
openreview.net
openreview.net
The Termination Critic
Anna Harutyunyan
Will Dabney
Diana L. Borsa
Nicolas Heess
Remi Munos
Doina Precup
In this work, we consider the problem of autonomously discovering behavioral abstractions, or options, for reinforcement learning agents. We… (see more) propose an algorithm that focuses on the termination function, as opposed to - as is common - the policy. The termination function is usually trained to optimize a control objective: an option ought to terminate if another has better value. We offer a different, information-theoretic perspective, and propose that terminations should focus instead on the compressibility of the option’s encoding - arguably a key reason for using abstractions.To achieve this algorithmically, we leverage the classical options framework, and learn the option transition model as a “critic” for the termination function. Using this model, we derive gradients that optimize the desired criteria. We show that the resulting options are non-trivial, intuitively meaningful, and useful for learning.
2019-02-26
ArXiv (preprint)
arxiv.org
arxiv.org
Predicting conversion to psychosis in clinical high risk patients using resting-state functional MRI features
Jolie Mcdonnell
W. Hord
Jenna Reinen
Pablo Polosecki
Irina Rish
Guillermo Cecchi
Recent progress in artificial intelligence provides researchers with a powerful set of machine learning tools for analyzing brain imaging da… (see more)ta. In this work, we explore a variety of classification algorithms and functional network features derived from resting-state fMRI data collected from clinical high-risk (prodromal schizophrenia) patients and controls, trying to identify features predictive of conversion to psychosis among a subset of CHR patients. While there are many existing studies suggesting that functional network features can be highly discriminative of schizophrenia when analyzing fMRI of patients suffering from the disease vs controls, few studies attempt to explore a similar approach to actual prediction of future psychosis development ahead of time, in the prodromal stage. Our preliminary results demonstrate the potential of fMRI functional network features to predict the conversion to psychosis in CHR patients. However, given the high variance of our results across different classifiers and subsets of data, a more extensive empirical investigation is required to reach more robust conclusions.
2019-02-16
Medical Imaging 2019: Biomedical Applications in Molecular, Structural, and Functional Imaging (published)
doi.org
Anytime Tail Averaging
Nicolas Le Roux
Tail averaging consists in averaging the last examples in a stream. Common techniques either have a memory requirement which grows with the … (see more)number of samples to average, are not available at every timestep or do not accomodate growing windows. We propose two techniques with a low constant memory cost that perform tail averaging with access to the average at every time step. We also show how one can improve the accuracy of that average at the cost of increased memory consumption.
2019-02-13
ArXiv (preprint)
arxiv.org
arxiv.org
Distributional reinforcement learning with linear function approximation
Marc Gendron-Bellemare
Nicolas Le Roux
Pablo Samuel Castro
Subhodeep Moitra
Despite many algorithmic advances, our theoretical understanding of practical distributional reinforcement learning methods remains limited.… (see more) One exception is Rowland et al. (2018)'s analysis of the C51 algorithm in terms of the Cramer distance, but their results only apply to the tabular setting and ignore C51's use of a softmax to produce normalized distributions. In this paper we adapt the Cramer distance to deal with arbitrary vectors. From it we derive a new distributional algorithm which is fully Cramer-based and can be combined to linear function approximation, with formal guarantees in the context of policy evaluation. In allowing the model's prediction to be any real vector, we lose the probabilistic interpretation behind the method, but otherwise maintain the appealing properties of distributional approaches. To the best of our knowledge, ours is the first proof of convergence of a distributional algorithm combined with function approximation. Perhaps surprisingly, our results provide evidence that Cramer-based distributional methods may perform worse than directly approximating the value function.
2019-02-08
ArXiv (preprint)
arxiv.org
arxiv.org
Distributional reinforcement learning with linear function approximation
Marc Gendron-Bellemare
Nicolas Le Roux
Pablo Samuel Castro
Subhodeep Moitra
Despite many algorithmic advances, our theoretical understanding of practical distributional reinforcement learning methods remains limited.… (see more) One exception is Rowland et al. (2018)'s analysis of the C51 algorithm in terms of the Cramer distance, but their results only apply to the tabular setting and ignore C51's use of a softmax to produce normalized distributions. In this paper we adapt the Cramer distance to deal with arbitrary vectors. From it we derive a new distributional algorithm which is fully Cramer-based and can be combined to linear function approximation, with formal guarantees in the context of policy evaluation. In allowing the model's prediction to be any real vector, we lose the probabilistic interpretation behind the method, but otherwise maintain the appealing properties of distributional approaches. To the best of our knowledge, ours is the first proof of convergence of a distributional algorithm combined with function approximation. Perhaps surprisingly, our results provide evidence that Cramer-based distributional methods may perform worse than directly approximating the value function.
2019-02-08
ArXiv (preprint)
arxiv.org
arxiv.org
Dendritic solutions to the credit assignment problem
Blake Richards
Timothy P. Lillicrap
2019-02-01
Current Opinion in Neurobiology (published)
doi.org
Equivalence of Equilibrium Propagation and Recurrent Backpropagation
Benjamin Scellier
Yoshua Bengio
Recurrent backpropagation and equilibrium propagation are supervised learning algorithms for fixed-point recurrent neural networks, which di… (see more)ffer in their second phase. In the first phase, both algorithms converge to a fixed point that corresponds to the configuration where the prediction is made. In the second phase, equilibrium propagation relaxes to another nearby fixed point corresponding to smaller prediction error, whereas recurrent backpropagation uses a side network to compute error derivatives iteratively. In this work, we establish a close connection between these two algorithms. We show that at every moment in the second phase, the temporal derivatives of the neural activities in equilibrium propagation are equal to the error derivatives computed iteratively by recurrent backpropagation in the side network. This work shows that it is not required to have a side network for the computation of error derivatives and supports the hypothesis that in biological neural networks, temporal derivatives of neural activities may code for error signals.
2019-02-01
Neural Computation (published)
doi.org
openreview.net