Filters








103,389 Hits in 4.4 sec

Robust Temporal Difference Learning for Critical Domains [article]

Richard Klima, Daan Bloembergen, Michael Kaisers, Karl Tuyls
2019 arXiv   pre-print
We present a new Q-function operator for temporal difference (TD) learning methods that explicitly encodes robustness against significant rare events (SRE) in critical domains.  ...  The operator, which we call the κ-operator, allows to learn a robust policy in a model-based fashion without actually observing the SRE.  ...  We are indebted to the anonymous reviewers of AAMAS 2019 for their valuable feedback.  ... 
arXiv:1901.08021v2 fatcat:nfmapvjklfgg3g72euvqprc4ti

Reinforcement Learning with Time-dependent Goals for Robotic Musicians [article]

Thilo Fryen, Manfred Eppe, Phuong D.H. Nguyen, Timo Gerkmann, Stefan Wermter
2020 arXiv   pre-print
In this paper, we address robotic musicianship by introducing a temporal extension to goal-conditioned reinforcement learning: Time-dependent goals.  ...  Reinforcement learning is a promising method to accomplish robotic control tasks.  ...  The authors gratefully acknowledge partial support from the German Research Foundation (DFG) under project Crossmodal Learning (TRR-169).  ... 
arXiv:2011.05715v1 fatcat:ed5lw46nmjaj3ccmdb3l4iqyn4

Audio-Visual Event Recognition through the lens of Adversary [article]

Juncheng B Li, Kaixin Ma, Shuhui Qu, Po-Yao Huang, Florian Metze
2020 arXiv   pre-print
2) How do different frequency/time domain features contribute to the robustness?  ...  As audio/visual classification models are widely deployed for sensitive tasks like content filtering at scale, it is critical to understand their robustness along with improving the accuracy.  ...  This is because ∞ is too potent as is shown in Fig. 2 (d) . 4 In the temporal domain, we can observe nearly uniform drop in performance despite of the different temporal masks, suggesting the temporal  ... 
arXiv:2011.07430v1 fatcat:ptkckp4y6vf5nnoovf2qzayeba

Recur, Attend or Convolve? On Whether Temporal Modeling Matters for Cross-Domain Robustness in Action Recognition [article]

Sofia Broomé, Ernest Pokropek, Boyu Li, Hedvig Kjellström
2022 arXiv   pre-print
In this article, we empirically study whether the choice of low-level temporal modeling has consequences for texture bias and cross-domain robustness.  ...  domains of Diving48 allowing for the investigation of texture bias for video models.  ...  Conclusions and discussion We have studied cross-domain robustness for three models that are principally different in terms of temporal modeling.  ... 
arXiv:2112.12175v3 fatcat:zae7zzdinzhujh63liwahwlesy

Joint Visual and Temporal Consistency for Unsupervised Domain Adaptive Person Re-Identification [article]

Jianing Li, Shiliang Zhang
2020 arXiv   pre-print
The two classification models are combined in a unified framework, which effectively leverages the unlabeled data for discriminative feature learning.  ...  For example, under unsupervised setting, our method outperforms recent unsupervised domain adaptive methods, which leverage more labels for training.  ...  Different from those works, we consider visual similarity and temporal consistency for feature learning.  ... 
arXiv:2007.10854v1 fatcat:mhr2ropgsrdfzgyiarn4mntv2i

Evaluation of Domain Generalization and Adaptation on Improving Model Robustness to Temporal Dataset Shift in Clinical Medicine [article]

Lin Lawrence Guo, Stephen R Pfohl, Jason Fries, Alistair Johnson, Jose Posada, Catherine Aftandilian, Nigam Shah, Lillian Sung
2021 medRxiv   pre-print
Algorithms that learn robust models by estimating invariant properties across time periods for domain generalization (DG) and unsupervised domain adaptation (UDA) might be suitable to proactively mitigate  ...  When compared with ERM[08-16], DG and UDA experiments failed to produce more robust models (range of AUROC difference, -0.003-0.050).  ...  manuscript or revising it critically for important intellectual content: AllFinal approval of version to be published: All Figure 1 .Figure 2 . 12 Data Splitting Procedure for baseline, domain generalization  ... 
doi:10.1101/2021.06.17.21259092 fatcat:xozhfnxda5bcznufbrs6mhthh4

Temporal Disentanglement of Representations for Improved Generalisation in Reinforcement Learning [article]

Mhairi Dunion, Trevor McInroe, Kevin Luck, Josiah Hanna, Stefano V. Albrecht
2022 arXiv   pre-print
To learn more robust representations, we introduce TEmporal Disentanglement (TED), a self-supervised auxiliary task that leads to disentangled representations using the sequential nature of RL observations  ...  This issue is intensified for image-based RL where a change in one variable, such as the background colour, can change many pixels in the image, and in turn can change all values in the agent's internal  ...  We introduce a self-supervised auxiliary task for learning disentangled representations for the robust encoding of images, which we call TEmporal Disentanglement (TED).  ... 
arXiv:2207.05480v1 fatcat:aiqpuygzzjgsfcsbjveqyzmdr4

Data-efficient Hindsight Off-policy Option Learning [article]

Markus Wulfmeier, Dushyant Rao, Roland Hafner, Thomas Lampe, Abbas Abdolmaleki, Tim Hertweck, Michael Neunert, Dhruva Tirumala, Noah Siegel, Nicolas Heess, Martin Riedmiller
2021 arXiv   pre-print
The approach outperforms existing option learning methods on common benchmarks.  ...  We introduce Hindsight Off-policy Options (HO2), a data-efficient option learning algorithm.  ...  We additionally like to acknowledge the support of the DeepMind robotics lab for infrastructure and engineering support.  ... 
arXiv:2007.15588v2 fatcat:ilf6pndw5zhnphh6drdgk2kc4q

Learning Robust Options

Daniel Mankowitz, Timothy Mann, Pierre-Luc Bacon, Doina Precup, Shie Mannor
2018 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
In this paper, we propose robust methods for learning temporally abstract actions, in the framework of options.  ...  We present a Robust Options Policy Iteration (ROPI) algorithm with convergence guarantees, which learns options that are robust to model uncertainty.  ...  It learns by minimizing the Temporal Difference (TD) loss. Typically, a separate DQN is trained to solve each task.  ... 
doi:10.1609/aaai.v32i1.12115 fatcat:tkejcsacszclbonmdhqiyexupm

Learning Robust Options [article]

Daniel J. Mankowitz, Timothy A. Mann, Pierre-Luc Bacon, Doina Precup, Shie Mannor
2018 arXiv   pre-print
In this paper, we propose robust methods for learning temporally abstract actions, in the framework of options.  ...  We present a Robust Options Policy Iteration (ROPI) algorithm with convergence guarantees, which learns options that are robust to model uncertainty.  ...  It learns by minimizing the Temporal Difference (TD) loss. Typically, a separate DQN is trained to solve each task.  ... 
arXiv:1802.03236v1 fatcat:24zvosmwjrg7dduowd74lhm72e

Video-Based Person Re-Identification: Methods, Datasets, and Deep Learning

2020 International Journal of Engineering and Advanced Technology  
Feature representation and metric learning are major issues for person re-identification.  ...  Current trends which include open re-identification systems, use of discriminative features and deep learning is marching towards new applications in security and surveillance, typically for tracking  ...  These methods integrate information extracted from the temporal domain with a spatial one, making it a case of Spatio-temporal matching.  ... 
doi:10.35940/ijeat.c6524.029320 fatcat:zbaiu3k7yncc3djk3onwf2jklq

Differencing based Self-supervised pretraining for Scene Change Detection [article]

Vijaya Raghavan T. Ramkumar, Elahe Arani, Bahram Zonooz
2022 arXiv   pre-print
On the other hand, transfer learning from large datasets induces domain shift.  ...  Our results also demonstrate the robustness of DSP to natural corruptions, distribution shift, and learning under limited labeled data.  ...  difference (changed) regions across temporal views to learn distinctive representation.  ... 
arXiv:2208.05838v1 fatcat:grdpx2ekfrcgzmn22vsggijn6e

Effects of enriched auditory experience on infants' speech perception during the first year of life

T. Christina Zhao, Patricia K. Kuhl
2016 Prospect: Quarterly Review of Comparative Education  
Results showed that a 1-month laboratory music intervention focusing on rhythm learning enhanced 9-month-old infants' neural processing not only for music but also for speech.  ...  Infants rapidly learn language in their home environments.  ...  the difference waves for the temporal regions of the cortex for the Intervention group and the Control group.  ... 
doi:10.1007/s11125-017-9397-6 fatcat:tmrx7h4bvvf4vouaa6skv7dmoy

Dynamic Layer Customization for Noise Robust Speech Emotion Recognition in Heterogeneous Condition Training [article]

Alex Wilf, Emily Mower Provost
2020 arXiv   pre-print
, we can use known noise conditions and domain adaptation algorithms to train systems that generalize well to unseen noise conditions.  ...  Robustness to environmental noise is important to creating automatic speech emotion recognition systems that are deployable in the real world.  ...  [8] for additional details. We use DLC to test different methods as unimodal acoustic feature encoders for HFFN, maintaining the original ordering HFFN requires for cross-utterance learning.  ... 
arXiv:2010.11226v1 fatcat:ytusez7cmvhhjgumyk6ykfb44q

Improving Video Generation for Multi-functional Applications [article]

Bernhard Kratzwald, Zhiwu Huang, Danda Pani Paudel, Acharya Dinesh, Luc Van Gool
2018 arXiv   pre-print
This is achieved by designing a robust one-stream video generation architecture with an extension of the state-of-the-art Wasserstein GAN framework that allows for better convergence.  ...  Our model can thus be trained to generate - and learn from - a broad set of videos with no restriction.  ...  Especially for video generation, it turns out to be much more challenging [7] as low frequencies also span the additional temporal domain.  ... 
arXiv:1711.11453v2 fatcat:tpacxtk3hvaqngqjpfwbrt42xe
« Previous Showing results 1 — 15 out of 103,389 results