49,858 Hits in 7.6 sec

Active Learning of Causal Structures with Deep Reinforcement Learning [article]

Amir Amirinezhad, Saber Salehkaleybar, Matin Hashemi
2020 arXiv   pre-print
The goal is to fully identify the causal structures with minimum number of interventions. We present the first deep reinforcement learning based solution for the problem of experiment design.  ...  We study the problem of experiment design to learn causal structures from interventional data.  ...  Unlike previous works on experiment design in the active setting, we will utilize deep reinforcement learning methods in order to find suitable Q function.  ... 
arXiv:2009.03009v1 fatcat:po7v5wv6dzfxti7oozixxnkc7y

Reinforcement Learning and its Connections with Neuroscience and Psychology [article]

Ajay Subramanian, Sharad Chitlangia, Veeky Baths
2021 arXiv   pre-print
In this paper, we comprehensively review a large number of findings in both neuroscience and psychology that evidence reinforcement learning as a promising candidate for modeling learning and decision  ...  These algorithms have outperformed humans in several tasks by learning from scratch, using only scalar rewards obtained through interaction with their environment.  ...  In formal terms of causal inference, with a structural causal model (SCM) of the environment, counterfactual scenarios can be simulated.  ... 
arXiv:2007.01099v5 fatcat:mjpkztlmqnfjba3dtcwqwmmlvu

Explainable Artificial Intelligence (xAI) Approaches and Deep Meta-Learning Models [chapter]

Evren Dağlarli
2020 Advances and Applications in Deep Learning  
However, we are still at the beginning of the way to understand these types of models. The forthcoming years are expected to be years in which the openness of deep learning models is discussed.  ...  , the activation functions, and the optimization algorithms.  ...  This causal structure learns the rules with its own internal deep learning method.  ... 
doi:10.5772/intechopen.92172 fatcat:sgmxtwloa5bbzb5sp7tpi75i3y

Towards Deep Symbolic Reinforcement Learning [article]

Marta Garnelo, Kai Arulkumaran, Murray Shanahan
2016 arXiv   pre-print
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  ...  In this paper, we propose an end-to-end reinforcement learning architecture comprising a neural back end and a symbolic front end with the potential to overcome each of these shortcomings.  ...  Acknowledgments We are grateful to Nvidia Corporation for the donation of a high-end GPU. Marta Garnelo is supported by an EPSRC doctoral training award. Author Contributions  ... 
arXiv:1609.05518v2 fatcat:vujv5k5ybzaitb72s62iyolvd4

Towards intervention-centric causal reasoning in learning agents [article]

Benjamin Lansdell
2020 arXiv   pre-print
This work shows how advances in deep reinforcement learning and meta-learning can provide intervention-centric causal learning in high-dimensional environments with a latent causal structure.  ...  space with a latent causal structure.  ...  The problem is that SCMs are hard to scale to high-dimensional state spaces, perhaps with a latent causal structure, and thus they are hard to combine with modern deep-learning-based methods (though see  ... 
arXiv:2005.12968v1 fatcat:pga3o7nxpzhh5pviodtrwfszfy

Temporal-Spatial Causal Interpretations for Vision-Based Reinforcement Learning [article]

Wenjie Shi, Gao Huang, Shiji Song, Cheng Wu
2021 arXiv   pre-print
Deep reinforcement learning (RL) agents are becoming increasingly proficient in a range of complex control tasks.  ...  TSCI model is applicable to recurrent agents and can be used to discover causal features with high efficiency once trained.  ...  terfactual analysis of saliency maps for deep reinforcement learning,” in [64] D. A.  ... 
arXiv:2112.03020v1 fatcat:qdypudi6djcanelgijzh3ifkqq

Explainability in reinforcement learning: perspective and position [article]

Agneza Krajna, Mario Brcic, Tomislav Lipic, Juraj Doncevic
2022 arXiv   pre-print
Reinforcement learning (RL) models increase the space of solvable problems with respect to other machine learning paradigms.  ...  Most articles dealing with explainability in artificial intelligence provide methods that concern supervised learning and there are very few articles dealing with this in the area of RL.  ...  Learning Policies, 2017. [42] Explainable Reinforcement Learning Through a Causal Lens, 2020. [55] Distal Explanations for Model-free Explainable Reinforcement Learning, 2020. [56] Deep Structural Causal  ... 
arXiv:2203.11547v1 fatcat:7zfi7f3i6bdgbhcjg7izvpnieq

Building Machines That Learn and Think Like People [article]

Brenden M. Lake, Tomer D. Ullman, Joshua B. Tenenbaum, Samuel J. Gershman
2016 arXiv   pre-print
We suggest concrete challenges and promising routes towards these goals that can combine the strengths of recent neural network advances with more structured cognitive models.  ...  Specifically, we argue that these machines should (a) build causal models of the world that support explanation and understanding, rather than merely solving pattern recognition problems; (b) ground learning  ...  Tom Schaul was very helpful in answering questions regarding the DQN learning curves and Frostbite scoring.  ... 
arXiv:1604.00289v3 fatcat:ph2rrwk2znb4dpb5nvcg54x2xi

A Survey of Deep Reinforcement Learning in Recommender Systems: A Systematic Review and Future Directions [article]

Xiaocong Chen, Lina Yao, Julian McAuley, Guanglin Zhou, Xianzhi Wang
2021 arXiv   pre-print
of the recent trends of deep reinforcement learning in recommender systems.  ...  In light of the emergence of deep reinforcement learning (DRL) in recommender systems research and several fruitful results in recent years, this survey aims to provide a timely and comprehensive overview  ...  area of deep reinforcement learning.  ... 
arXiv:2109.03540v2 fatcat:5gwrbfcj3rc7jfkd54eseck5ga

Schema Networks: Zero-shot Transfer with a Generative Causal Model of Intuitive Physics [article]

Ken Kansky, Tom Silver, David A. Mély, Mohamed Eldawy, Miguel Lázaro-Gredilla, Xinghua Lou, Nimrod Dorfman, Szymon Sidor, Scott Phoenix, Dileep George
2017 arXiv   pre-print
The recent adaptation of deep neural network-based methods to reinforcement learning and planning domains has yielded remarkable progress on individual tasks.  ...  The richly structured architecture of the Schema Network can learn the dynamics of an environment directly from data.  ...  Acknowledgements Special thanks to Eric Purdy and Ramki Gummadi for useful insights and discussions during the preparation of this work.  ... 
arXiv:1706.04317v2 fatcat:yrzp5yp2gvfflinghomly7ajsm

Deep Brain Stimulation Programming 2.0: Future Perspectives for Target Identification and Adaptive Closed Loop Stimulation

Franz Hell, Carla Palleis, Jan H. Mehrkens, Thomas Koeglsperger, Kai Bötzel
2019 Frontiers in Neurology  
Deep brain stimulation has developed into an established treatment for movement disorders and is being actively investigated for numerous other neurological as well as psychiatric disorders.  ...  Improving the targeting of anatomical and functional networks involved in the generation of pathological neural activity will improve the clinical DBS effect and limit side-effects.  ...  In reinforcement learning, an agent, in this case the DBS stimulation controller interacts with an uncertain environment, i.e., stimulating a mixture of neural structures with certain stimulation parameters  ... 
doi:10.3389/fneur.2019.00314 pmid:31001196 pmcid:PMC6456744 fatcat:gr2azomkijaeviu2hay6f4slyy

Autonomous development and learning in artificial intelligence and robotics: Scaling up deep learning to human-like learning

Pierre-Yves Oudeyer
2017 Behavioral and Brain Sciences  
These mechanisms guide exploration and autonomous choice of goals, and integrating them with deep learning opens stimulating perspectives.  ...  Autonomous lifelong development and learning are fundamental capabilities of humans, differentiating them from current deep learning systems.  ...  Recent work in deep reinforcement learning has included some of these mechanisms to solve difficult reinforcement learning problems, with rare or deceptive rewards (Bellemare et al. 2016; Kulkarni et  ... 
doi:10.1017/s0140525x17000243 pmid:29342696 fatcat:lgzpqnpl4rheteun7cquan7f44

Building machines that learn and think like people

Brenden M. Lake, Tomer D. Ullman, Joshua B. Tenenbaum, Samuel J. Gershman
2016 Behavioral and Brain Sciences  
We suggest concrete challenges and promising routes toward these goals that can combine the strengths of recent neural network advances with more structured cognitive models.  ...  Specifically, we argue that these machines should (1) build causal models of the world that support explanation and understanding, rather than merely solving pattern recognition problems; (2) ground learning  ...  Tom Schaul was very helpful in answering questions regarding the DQN learning curves and Frostbite scoring.  ... 
doi:10.1017/s0140525x16001837 pmid:27881212 fatcat:3fjriprksbhaxpqdcydrhmcjqm

Causal Reasoning from Meta-reinforcement Learning [article]

Ishita Dasgupta, Jane Wang, Silvia Chiappa, Jovana Mitrovic, Pedro Ortega, David Raposo, Edward Hughes, Peter Battaglia, Matthew Botvinick, Zeb Kurth-Nelson
2019 arXiv   pre-print
We train a recurrent network with model-free reinforcement learning to solve a range of problems that each contain causal structure.  ...  This work also offers new strategies for structured exploration in reinforcement learning, by providing agents with the ability to perform -- and interpret -- experiments.  ...  reinforcement learning (RL).  ... 
arXiv:1901.08162v1 fatcat:ie4ifxdojncdrn3axnh74c2ksi

Prospective Learning: Back to the Future [article]

Joshua T. Vogelstein, Timothy Verstynen, Konrad P. Kording, Leyla Isik, John W. Krakauer, Ralph Etienne-Cummings, Elizabeth L. Ogburn, Carey E. Priebe, Randal Burns, Kwame Kutten, James J. Knierim, James B. Potash (+51 others)
2022 arXiv   pre-print
Causal estimation enables learning the structure of relations that guide choosing actions for specific outcomes, even when the specific action-outcome contingencies have never been observed before.  ...  We call this 'retrospective learning'. For example, an intelligence may see a set of pictures of objects, along with their names, and learn to name them.  ...  Human-level control through deep reinforcement learn- ing.  ... 
arXiv:2201.07372v1 fatcat:6qktqmoffvd4zl37ubd4wo2f2m
« Previous Showing results 1 — 15 out of 49,858 results