Filters








6 Hits in 4.8 sec

Reconstruction of Fragmented Trajectories of Collective Motion using Hadamard Deep Autoencoders [article]

Kelum Gajamannage, Yonggi Park, Randy Paffenroth, Anura P. Jayasumana
2021 arXiv   pre-print
The trajectories of the agents practicing collective motion is low-rank due to mutual interactions and dependencies between the agents that we utilize as the underlying pattern that our Hadamard deep autoencoder  ...  The performance of our HDA is compared with that of a low-rank matrix completion scheme in the context of fragmented trajectory reconstruction.  ...  We presented a nonlinear machine learning technique for fragmented trajectory reconstruction in this paper which is made by incorporating an indicator matrix into a regular deep autoencoder using Hadamard  ... 
arXiv:2110.10428v1 fatcat:akfo4kjkwrgu3bvcn5b35dszru

"Forget" the Forget Gate: Estimating Anomalies in Videos Using Self-contained Long Short-Term Memory Networks [chapter]

Habtamu Fanta, Zhiwen Shao, Lizhuang Ma
2020 Lecture Notes in Computer Science  
Abnormal event detection is a challenging task that requires effectively handling intricate features of appearance and motion.  ...  The forget gate mitigates participation of previous hidden state for computation of cell state prioritizing current input.  ...  [33] , trajectories having spatial proximity manifesting related motion patterns are classified and used to identify outliers.  ... 
doi:10.1007/978-3-030-61864-3_15 fatcat:yf4nmh4tfnb7ln6m7fgsoyrpju

Knowledge Augmented Machine Learning with Applications in Autonomous Driving: A Survey [article]

Julian Wörmann, Daniel Bogdoll, Etienne Bührle, Han Chen, Evaristus Fuh Chuo, Kostadin Cvejoski, Ludger van Elst, Tobias Gleißner, Philip Gottschall, Stefan Griesche, Christian Hellert, Christian Hesels (+34 others)
2022 arXiv   pre-print
As a consequence, the reliable use of these models, especially in safety-critical applications, is a huge challenge.  ...  However, the subsequent application of these models often involves scenarios that are inadequately represented in the data used for training.  ...  need of collecting experimental data.  ... 
arXiv:2205.04712v1 fatcat:u2bgxr2ctnfdjcdbruzrtjwot4

Quantum computing enhanced machine learning for physico-chemical applications [article]

Manas Sajjan, Junxu Li, Raja Selvarajan, Shree Hari Sureshbabu, Sumit Suresh Kale, Rishabh Gupta, Sabre Kais
2021 arXiv   pre-print
We shall not only present a brief overview of the well-known techniques but also highlight their learning strategies using statistical physical insight.  ...  In the recent years, it is safe to conclude that ML and its close cousin deep learning (DL) have ushered unprecedented developments in all areas of physical sciences especially chemistry.  ...  illustrated in Fig. 14 (b) Autoencoders A typical autoencoder is a type of neural network which is used to generate useful representations of the input data, to be used for unsupervised learning.  ... 
arXiv:2111.00851v1 fatcat:i2caiglszvbufbyfmf3cwkcduu

Robust Methods for the Automatic Quantification and Prediction of Affect in Spoken Interactions

Zakaria Aldeneh, University, My
2021
This dissertation explores methodologies for improving the robustness of the automatic recognition of emotional expression from speech by addressing the impacts of these factors on various aspects of the  ...  For addressing data sparsity, we investigate two methods that enable us to learn robust embeddings, which highlight the differences that exist between neutral speech and emotionally expressive speech,  ...  trajectories.  ... 
doi:10.7302/29 fatcat:5drofg5jrjb57dar2bohimrl6m

Robust subspace learning for static and dynamic affect and behaviour modelling

Christos Georgakis, Maja Pantic, Engineering And Physical Sciences Research Council, European Commission
2017
Having identified a gap in the literature which is the lack of data containing annotations of social attitudes in conti [...]  ...  Inspired by the well-documented importance of the temporal aspect in perceiving affect and behavior, we direct the bulk of our research efforts into continuous-time modeling of dimensional affect and social  ...  as an unsupervised learning algorithm in distinguishing motion trajectories corresponding to different objects/persons.  ... 
doi:10.25560/52432 fatcat:5vuiijjh6nc3bj75jlx3jifwlq