Filters








1,298 Hits in 4.9 sec

Understanding Human Behaviors in Crowds by Imitating the Decision-Making Process

Haosheng Zou, Hang Su, Shihong Song, Jun Zhu
2018 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Crowd behavior understanding is crucial yet challenging across a wide range of applications, since crowd behavior is inherently determined by a sequential decision-making process based on various factors  ...  Different factors of human decision making are disentangled with mutual information maximization, with the process modeled by collision avoidance regularization and Social-Aware LSTMs.  ...  , presumably due to the difficulty of training recurrent neural network generators with GANs (Metz et al. 2017) .  ... 
doi:10.1609/aaai.v32i1.12316 fatcat:xfgbgw24ofexvchqgzeio6k3ou

Understanding Human Behaviors in Crowds by Imitating the Decision-Making Process [article]

Haosheng Zou, Hang Su, Shihong Song, Jun Zhu
2018 arXiv   pre-print
Crowd behavior understanding is crucial yet challenging across a wide range of applications, since crowd behavior is inherently determined by a sequential decision-making process based on various factors  ...  Different factors of human decision making are disentangled with mutual information maximization, with the process modeled by collision avoidance regularization and Social-Aware LSTMs.  ...  , presumably due to the difficulty of training recurrent neural network generators with GANs (Metz et al. 2017) .  ... 
arXiv:1801.08391v1 fatcat:qhc2xe75cfgghfq2vzv6s7pgpy

Grouptron: Dynamic Multi-Scale Graph Convolutional Networks for Group-Aware Dense Crowd Trajectory Forecasting [article]

Rui Zhou, Hongyu Zhou, Huidong Gao, Masayoshi Tomizuka, Jiachen Li, Zhuo Xu
2022 arXiv   pre-print
understanding and representation of the scenes.  ...  It then uses graph neural networks to encode dynamics at different scales and aggregate the embeddings for trajectory prediction.  ...  ., Gaussian Process Regression (GPR) [10] , inverse reinforcement learning (IRL) [11] , and recurrent neural networks (RNNs) [1] , [12] , [13] .  ... 
arXiv:2109.14128v3 fatcat:pe634vtaujabpjlygk5gvvsft4

Parameter Compression of Recurrent Neural Networks and Degradation of Short-term Memory [article]

Jonathan A. Cox
2017 arXiv   pre-print
recurrent neural networks (RNN).  ...  In this way, we demonstrate that the effect of compression of recurrent parameters is dependent on the degree of temporal coherence present in the data and task.  ...  In contrast to feed-forward networks, the recurrent neural network is also unfolded into a deep network in time, with shared recurrent weights at every time step.  ... 
arXiv:1612.00891v2 fatcat:cbfec2yfffhbxj7n36rzddqwbi

Story Generation from Sequence of Independent Short Descriptions [article]

Parag Jain, Priyanka Agrawal, Abhijit Mishra, Mohak Sukhwani, Anirban Laha, Karthik Sankaranarayanan
2017 arXiv   pre-print
We then implement a deep recurrent neural network (RNN) architecture that encodes sequence of variable length input descriptions to corresponding latent representations and decodes them to produce well  ...  This paper introduces and addresses the task of coherent story generation from independent descriptions, describing a scene or an event.  ...  We leverage 2 architecture i.e., deep recurrent neural network ( ) with a ention mechanism for the problem of coherent story generation.  ... 
arXiv:1707.05501v2 fatcat:oaa5ficdpvaefoibpeaa63lje4

Why visual attention and awareness are different

Victor A.F. Lamme
2003 Trends in Cognitive Sciences  
Conscious stimuli have reached a level of processing beyond initial feature detection, where at least an initial coherent perceptual interpretation of the scene is achieved.  ...  Visual awareness requires recurrent processing What remains to be found, then, is a similar core understanding of phenomenal experience.  ...  Questions for future research † Why does recurrent neural activity generate phenomenal experience whereas neural activity per se seems insufficient?  ... 
doi:10.1016/s1364-6613(02)00013-x pmid:12517353 fatcat:w6teycxgr5dyblnzcxyt7oqjte

HGCN-GJS: Hierarchical Graph Convolutional Network with Groupwise Joint Sampling for Trajectory Prediction [article]

Yuying Chen, Congcong Liu, Bertram E. Shi, Ming Liu
2020 arXiv   pre-print
In this work, we propose a hierarchical graph convolutional network, HGCN-GJS, for trajectory prediction which well leverages group level interactions within the crowd.  ...  Fully investigating the social interactions within the crowd is crucial for accurate pedestrian trajectory prediction.  ...  Recent works using recurrent neural networks (RNNs), like long short-term memory networks (LSTMs), have achieved great successes in trajectory prediction tasks (Alahi et al. 2016; Su et al. 2017; Hasan  ... 
arXiv:2009.07140v1 fatcat:eez5qs6mezhjdf7tqlg3dtezii

Convolutional Neural Network for Trajectory Prediction [article]

Nishant Nikhil, Brendan Tran Morris
2018 arXiv   pre-print
In this work, we propose a convolutional neural network (CNN) based human trajectory prediction approach.  ...  Predicting trajectories of pedestrians is quintessential for autonomous robots which share the same environment with humans.  ...  Conclusions We present a convolutional architecture based neural network model for trajectory prediction.  ... 
arXiv:1809.00696v2 fatcat:2iyaoqmscjbvrfommomtcommxa

Convolutional Neural Network for Trajectory Prediction [chapter]

Nishant Nikhil, Brendan Tran Morris
2019 Lecture Notes in Computer Science  
In this work, we propose a convolutional neural network (CNN) based human trajectory prediction approach.  ...  Predicting trajectories of pedestrians is quintessential for autonomous robots which share the same environment with humans.  ...  Conclusions We present a convolutional architecture based neural network model for trajectory prediction.  ... 
doi:10.1007/978-3-030-11015-4_16 fatcat:lrulrukhffbu3oebqqnog44txq

Separate neural definitions of visual consciousness and visual attention; a case for phenomenal awareness

V.A.F Lamme
2004 Neural Networks  
In the model proposed here, visual attention is defined as a convolution of sensori-motor processing with memory. Consciousness, however, is generated by recurrent activity between cortical areas.  ...  From these experiments clearly separate neural definitions of visual attention and visual consciousness emerge.  ...  V.A.F.Lamme / Neural Networks 17 (2004) 861-872  ... 
doi:10.1016/j.neunet.2004.02.005 pmid:15288903 fatcat:nplcyonskrgoxakye7obobgz24

Review of Pedestrian Trajectory Prediction Methods: Comparing Deep Learning and Knowledge-based Approaches [article]

Raphael Korbmacher, Antoine Tordeux
2021 arXiv   pre-print
The topology of the scene and the interactions between the pedestrians are just some of them.  ...  In crowd scenarios, predicting trajectories of pedestrians is a complex and challenging task depending on many external factors.  ...  neural networks with more than two hidden layers (deep neural networks) [32] .  ... 
arXiv:2111.06740v1 fatcat:lstxi2qrhrhdtlpm47x4vpinry

Pedestrian Path Forecasting in Crowd

Yuke Li
2017 Proceedings of the 2017 ACM on Multimedia Conference - MM '17  
Predicting the walking path of a pedestrian in crowds is a pivotal step towards understanding his/her behavior.  ...  We evaluate our approach on three large benchmark datasets, and show that it introduces large margin improvements with respect to recent works in the literature, both in short and long-term forecasting  ...  Recently, Deep Neural Networks (DNNs) [9] have shown to achieve cutting-edge performance in several vision tasks related to visual understanding [5, 12, 13, 23, 26] , action recognition [25, 27, 43  ... 
doi:10.1145/3123266.3123287 dblp:conf/mm/Li17 fatcat:exnzfpj6yrhyzd4ekbjgztucii

Deep Tracking on the Move: Learning to Track the World from a Moving Vehicle using Recurrent Neural Networks [article]

Julie Dequaire, Dushyant Rao, Peter Ondruska, Dominic Wang, Ingmar Posner
2017 arXiv   pre-print
Inspired by the recently presented DeepTracking approach [Ondruska, 2016], we employ a recurrent neural network (RNN) to capture the temporal evolution of the state of the environment, and propose to use  ...  This paper presents an end-to-end approach for tracking static and dynamic objects for an autonomous vehicle driving through crowded urban environments.  ...  In the static scenario, the dynamics of the scene as viewed from the world frame are coherent with that viewed from the local sensor frame.  ... 
arXiv:1609.09365v3 fatcat:55hw7b4xqjczhjpnqgpspxejli

Towards Task Understanding in Visual Settings [article]

Sebastin Santy, Wazeer Zulfikar, Rishabh Mehrotra, Emine Yilmaz
2018 arXiv   pre-print
We leverage insights from real world task understanding systems, and propose a framework composed of convolutional neural networks, and an external hierarchical task ontology to produce task descriptions  ...  While most existing image captioning methods excel in producing natural language descriptions of visual scenes involving human tasks, there is often the need for an understanding of the exact task being  ...  NeuralTalk2 uses convolutional and recurrent neural networks in multimodal space to generate image descriptions. im2txt is similar to NeuralTalk2 with a better classifier.  ... 
arXiv:1811.11833v1 fatcat:5p7bbkztszdh3n25m5mysm3faq

Towards Task Understanding in Visual Settings

Sebastin Santy, Wazeer Zulfikar, Rishabh Mehrotra, Emine Yilmaz
2019 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
We leverage insights from real world task understanding systems, and propose a framework composed of convolutional neural networks, and an external hierarchical task ontology to produce task descriptions  ...  While most existing image captioning methods excel in producing natural language descriptions of visual scenes involving human tasks, there is often the need for an understanding of the exact task being  ...  Neu-ralTalk2 uses convolutional and recurrent neural networks in multimodal space to generate image descriptions. im2txt is similar to NeuralTalk2 with a better classifier.  ... 
doi:10.1609/aaai.v33i01.330110027 fatcat:unfiiap7mreebk6dr5a4vbyk4a
« Previous Showing results 1 — 15 out of 1,298 results