Filters








29,203 Hits in 4.7 sec

Building Generalizable Agents with a Realistic and Rich 3D Environment [article]

Yi Wu, Yuxin Wu, Georgia Gkioxari, Yuandong Tian
2018 arXiv   pre-print
Using a subset of houses in House3D, we show that reinforcement learning agents trained with an enhancement of different levels of augmentations perform much better in unseen environments than our baselines  ...  The diversity in House3D opens the door towards scene-level augmentation, while the label-rich nature of House3D enables us to inject pixel- & task-level augmentations such as domain randomization (Toubin  ...  A number of recent works also use deep reinforcement learning for navigation in simulated 3D scenes.  ... 
arXiv:1801.02209v2 fatcat:27gzhfbhbbeknbwmeie5sg5nnm

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.  ...  Jimenez Rezende, “Towards interpretable reinforcement learning via attention in atari agents,” in 2019 IEEE/CVF International Confer- using attention augmented agents,” in Advances  ... 
arXiv:2112.03020v1 fatcat:qdypudi6djcanelgijzh3ifkqq

Show, attend and interact: Perceivable human-robot social interaction through neural attention Q-network

Ahmed Hussain Qureshi, Yutaka Nakamura, Yuichiro Yoshikawa, Hiroshi Ishiguro
2017 2017 IEEE International Conference on Robotics and Automation (ICRA)  
We introduce the Multimodal Deep Attention Recurrent Q-Network using which the robot exhibits human-like social interaction skills after 14 days of interacting with people in an uncontrolled real world  ...  Each and every day during the 14 days, the system gathered robot interaction experiences with people through a hit-and-trial method and then trained the MDARQN on these experiences using end-to-end reinforcement  ...  Recent advancements in machine learning has combined deep learning with reinforcement learning and has led to the development of Deep Q-Network (DQN) [4] .  ... 
doi:10.1109/icra.2017.7989193 dblp:conf/icra/QureshiNYI17 fatcat:op4o2nah35hm5luy2plaavx2f4

Interactive Reinforcement Learning for Object Grounding via Self-Talking [article]

Yan Zhu, Shaoting Zhang, Dimitris Metaxas
2017 arXiv   pre-print
During interactive training, both agents are reinforced by the guidance from a common reward function.  ...  However, we observe language drifting problem during training and propose to use reward engineering to improve the interpretability for the generated conversations.  ...  The question generator and the answer models are collectively tuned by a common reward function using reinforcement learning.  ... 
arXiv:1712.00576v1 fatcat:gtq6b2dbc5b7pdrydo6s5j7bku

Autonomous Industrial Management via Reinforcement Learning: Self-Learning Agents for Decision-Making – A Review [article]

Leonardo A. Espinosa Leal, Magnus Westerlund, Anthony Chapman
2019 arXiv   pre-print
We discuss using different types of extended realities, such as digital twins, to train reinforcement learning agents to learn specific tasks through generalization.  ...  Once generalization is achieved, we discuss how these can be used to develop self-learning agents.  ...  Assuming that the cortical conductor theory interpretation is correct [18] , we can reduce an autonomous system into four main concepts that feeds an attention layer, learning, reasoning, control, and  ... 
arXiv:1910.08942v1 fatcat:wbpy3iijhbfwxeatiy4ztt5f2m

Interactive Language Acquisition with One-shot Visual Concept Learning through a Conversational Game

Haichao Zhang, Haonan Yu, Wei Xu
2018 Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)  
Building intelligent agents that can communicate with and learn from humans in natural language is of great value.  ...  The agent trained with this approach is able to actively acquire information by asking questions about novel objects and use the justlearned knowledge in subsequent conversations in a one-shot fashion.  ...  The learner can ask about the new class and use the interpreter to extract useful information from the teacher's sentence via word-level attention η and content importance g mem jointly.  ... 
doi:10.18653/v1/p18-1243 dblp:conf/acl/XuYZ18 fatcat:tq2ddmukovgqnfevi4payy36qe

Memory-Augmented Reinforcement Learning for Image-Goal Navigation [article]

Lina Mezghani, Sainbayar Sukhbaatar, Thibaut Lavril, Oleksandr Maksymets, Dhruv Batra, Piotr Bojanowski, Karteek Alahari
2022 arXiv   pre-print
Our method is based on an attention-based end-to-end model that leverages an episodic memory to learn to navigate.  ...  In this work, we present a memory-augmented approach for image-goal navigation.  ...  Given the difficulty to interpret raw RGB observations, the agent needs to learn a more informative representation than pixels.  ... 
arXiv:2101.05181v4 fatcat:v5mva4caxndu7k77iro5slgnhu

Augmented Replay Memory in Reinforcement Learning With Continuous Control [article]

Mirza Ramicic, Andrea Bonarini
2019 arXiv   pre-print
Online reinforcement learning agents are currently able to process an increasing amount of data by converting it into a higher order value functions.  ...  The similar dynamics are implemented by a proposed augmented memory replay AMR capable of optimizing the replay of the experiences from the agent's memory structure by altering or augmenting their relevance  ...  In the experiments reported in this paper, the augmentation dynamics are evolved over generations of learning agents performing reinforcement learning tasks in various environments.  ... 
arXiv:1912.12719v1 fatcat:ut5iv2ghsrdmtnwdzzcr6bbtnq

Speaker-Follower Models for Vision-and-Language Navigation [article]

Daniel Fried, Ronghang Hu, Volkan Cirik, Anna Rohrbach, Jacob Andreas, Louis-Philippe Morency, Taylor Berg-Kirkpatrick, Kate Saenko, Dan Klein, Trevor Darrell
2018 arXiv   pre-print
We use this speaker model to (1) synthesize new instructions for data augmentation and to (2) implement pragmatic reasoning, which evaluates how well candidate action sequences explain an instruction.  ...  reasoning process using generic sequence models.  ...  This work was partially supported by US DoD and DARPA XAI and D3M, NSF awards IIS-1833355, Oculus VR, and the Berkeley Artificial Intelligence Research (BAIR) Lab.  ... 
arXiv:1806.02724v2 fatcat:t6g46qwawbgd3b4n3snyvzhywe

TarMAC: Targeted Multi-Agent Communication [article]

Abhishek Das, Théophile Gervet, Joshua Romoff, Dhruv Batra, Devi Parikh, Michael Rabbat, Joelle Pineau
2020 arXiv   pre-print
We propose a targeted communication architecture for multi-agent reinforcement learning, where agents learn both what messages to send and whom to address them to while performing cooperative tasks in  ...  Moreover, we show that the targeted communication strategies learned by agents are interpretable and intuitive.  ...  Multi-Agent Reinforcement Learning (MARL).  ... 
arXiv:1810.11187v2 fatcat:gi4oogacaraizpdbce5qmittl4

Future Trends for Human-AI Collaboration: A Comprehensive Taxonomy of AI/AGI Using Multiple Intelligences and Learning Styles [article]

Andrzej Cichocki, Alexander P. Kuleshov
2020 arXiv   pre-print
Multi-agent systems such as these can be used to solve problems that would be difficult to solve by any individual intelligent agent.  ...  Key words: Artificial General Intelligence (AGI), multiple intelligences, learning styles, physical intelligence, emotional intelligence, social intelligence, attentional intelligence, moral-ethical intelligence  ...  In current AI systems, we extensively use six basic ways of learning: Supervised learning, unsupervised learning, semi-supervised learning, reinforcement learning, ensemble learning and deep learning (  ... 
arXiv:2008.04793v4 fatcat:4l4wxa3bwnhlfbkfp2i63uwaou

Free-Lunch Saliency via Attention in Atari Agents [article]

Dmitry Nikulin, Anastasia Ianina, Vladimir Aliev, Sergey Nikolenko
2019 arXiv   pre-print
learning agents.  ...  We propose a new approach to visualize saliency maps for deep neural network models and apply it to deep reinforcement learning agents trained on Atari environments.  ...  Over the last decade, reinforcement learning has shifted towards deep reinforcement learning, where policies and/or models of the environment are modeled with deep neural networks, with impressive results  ... 
arXiv:1908.02511v2 fatcat:7iwnlsix7fgn7g46pl66a7pb3e

Interactive Language Acquisition with One-shot Visual Concept Learning through a Conversational Game [article]

Haichao Zhang, Haonan Yu, Wei Xu
2018 arXiv   pre-print
The agent trained with this approach is able to actively acquire information by asking questions about novel objects and use the just-learned knowledge in subsequent conversations in a one-shot fashion  ...  Building intelligent agents that can communicate with and learn from humans in natural language is of great value.  ...  The learner can ask about the new class and use the interpreter to extract useful information from the teacher's sentence via word-level attention η and content importance g mem jointly.  ... 
arXiv:1805.00462v1 fatcat:exp4tpepobb2tf2uvydgqtx4je

Reinforcement Learning with Attention that Works: A Self-Supervised Approach [article]

Anthony Manchin, Ehsan Abbasnejad, Anton van den Hengel
2019 arXiv   pre-print
However previous attempts at integrating attention with reinforcement learning have failed to produce significant improvements.  ...  We propose the first combination of self attention and reinforcement learning that is capable of producing significant improvements, including new state of the art results in the Arcade Learning Environment  ...  attention with reinforcement learning.  ... 
arXiv:1904.03367v1 fatcat:5qlgijugpjfd3p6wzvb4m27u6y

Reinforcement Learning with Human Teachers: Understanding How People Want to Teach Robots

Andrea L. Thomaz, Guy Hoffman, Cynthia Breazeal
2006 ROMAN 2006 - The 15th IEEE International Symposium on Robot and Human Interactive Communication  
While Reinforcement Learning (RL) is not traditionally designed for interactive supervisory input from a human teacher, several works in both robot and software agents have adapted it for human input by  ...  We report three main observations on how people administer feedback when teaching a robot a task through Reinforcement Learning: (a) they use the reward channel not only for feedback, but also for future-directed  ...  Several examples of agents that learn interactively with a human teacher are based on Reinforcement Learning (RL).  ... 
doi:10.1109/roman.2006.314459 dblp:conf/ro-man/ThomazHB06 fatcat:hmssozyf7rdk7ghuzzi6vaiu34
« Previous Showing results 1 — 15 out of 29,203 results