45,349 Hits in 8.4 sec

Online Fast Adaptation and Knowledge Accumulation: a New Approach to Continual Learning [article]

Massimo Caccia, Pau Rodriguez, Oleksiy Ostapenko, Fabrice Normandin, Min Lin, Lucas Caccia, Issam Laradji, Irina Rish, Alexandre Lacoste, David Vazquez, Laurent Charlin
2021 arXiv   pre-print
We empirically show that Continual-MAML is better suited to the new scenario than the aforementioned methodologies, as well as standard continual learning and meta-learning approaches.  ...  Continual learning studies agents that learn from streams of tasks without forgetting previous ones while adapting to new ones.  ...  Acknowledgments and Disclosure of Funding Laurent Charlin is supported through a CIFAR AI Chair and grants from NSERC, CIFAR, IVADO, Samsung, and Google.  ... 
arXiv:2003.05856v3 fatcat:atm3wjdxvbd6xgjrbepkafdqza

Online Fast Adaptation and Knowledge Accumulation (OSAKA): a New Approach to Continual Learning

Massimo Caccia, Pau Rodríguez, Oleksiy Ostapenko, Fabrice Normandin, Min Lin, Lucas Page-Caccia, Issam Hadj Laradji, Irina Rish, Alexandre Lacoste, David Vázquez, Laurent Charlin
2020 Neural Information Processing Systems  
We show in an empirical study that Continual-MAML is better suited to the new scenario than the aforementioned methodologies including standard continual learning and meta-learning approaches.  ...  Continual learning agents experience a stream of (related) tasks. The main challenge is that the agent must not forget previous tasks and also adapt to novel tasks in the stream.  ...  Acknowledgments and Disclosure of Funding Laurent Charlin is supported through a CIFAR AI Chair and grants from NSERC, CIFAR, IVADO, Samsung, and Google.  ... 
dblp:conf/nips/CacciaRONLPLRLV20 fatcat:enj2ptqz6zhhvj6imbl7h6cb4y

Learn Continually, Generalize Rapidly: Lifelong Knowledge Accumulation for Few-shot Learning [article]

Xisen Jin, Bill Yuchen Lin, Mohammad Rostami, Xiang Ren
2021 arXiv   pre-print
We examine how the generalization ability is affected in the continual learning setup, evaluate a number of continual learning algorithms, and propose a novel regularized adapter generation approach.  ...  The ability to continuously expand knowledge over time and utilize it to rapidly generalize to new tasks is a key feature of human linguistic intelligence.  ...  Motivated by this observation, we study whether NLP models could accumulate generalizable knowledge continuously over a sequence of tasks and learn to generalize to new tasks rapidly (i.e., with few examples  ... 
arXiv:2104.08808v3 fatcat:q4gs4qw6uzgbda2hkv7oju53di

Transforming theological education is not the accumulation of knowledge, but the development of consciousness

Johannes J. Knoetze
2020 Verbum et Ecclesia  
to continue as faculties of theology (and religion).  ...  Currently there is a new realisation that these faculties need to accommodate Independent and Pentecostal churches if the church wants to further influence South African society and the faculties want  ...  Data availability statement Data sharing is not applicable to this article as no new data were created or analysed in this study.  ... 
doi:10.4102/ve.v41i1.2075 fatcat:lyssb73t4ngxpdk4beg5r3gkma

Ubiquitous media agents: a framework for managing personally accumulated multimedia files

Liu Wenyin, Zheng Chen, Fan Lin, Hongjiang Zhang, Wei-Ying Ma
2003 Multimedia Systems  
The media agents learn from the user's interaction records to refine the semantic indices and to model user intentions and preferences.  ...  Various algorithms can be used to implement the framework, and a few of them are described in this paper.  ...  index (or personal index), and (3) the learning algorithms: learning and refining the semantics based on user logs and learning to predict user actions given the knowledge base and information typed by  ... 
doi:10.1007/s00530-003-0085-4 fatcat:tfjtipsy5rfzdfpsevxc4f5lei

Fast bacterial growth reduces antibiotic accumulation and efficacy

Urszula Łapińska, Margaritis Voliotis, Ka Kiu Lee, Adrian Campey, M Rhia L Stone, Brandon Tuck, Wanida Phetsang, Bing Zhang, Krasimira Tsaneva-Atanasova, Mark AT Blaskovich, Stefano Pagliara
2022 eLife  
We used this new knowledge to eradicate variants that displayed low antibiotic accumulation through the chemical manipulation of their outer membrane inspiring new avenues to overcome current antibiotic  ...  This approach allowed us to rapidly identify phenotypic variants that avoid antibiotic accumulation within populations of Escherichia coli, Pseudomonas aeruginosa, Burkholderia cenocepacia, and Staphylococcus  ...  Acknowledgements U.L. was supported through a BBSRC responsive mode grant (BB/V008021/1), an MRC Proximity to Discovery EXCITEME2 grant (MCPC17189) and an award from the Gordon and Betty Moore Foundation  ... 
doi:10.7554/elife.74062 pmid:35670099 pmcid:PMC9173744 fatcat:3j7kifosh5flnhouxxhq2usrwy

Auditory Streaming as an Online Classification Process with Evidence Accumulation

Dana Barniv, Israel Nelken, Joel Snyder
2015 PLoS ONE  
Evidence for segregation is accumulated during the integration phase and vice versa; a switch to the opposite percept occurs stochastically based on this evidence.  ...  To account together for bistability and positive correlation between subsequent durations, we suggest that streaming is a consequence of an evidence accumulation process.  ...  We thank Misha Tsodyks for his support and useful comments. Author Contributions  ... 
doi:10.1371/journal.pone.0144788 pmid:26671774 pmcid:PMC4699212 fatcat:3uaqhe7ldfb3dasvqhtz32jfeu

What is value—accumulated reward or evidence?

Karl Friston, Rick Adams, Read Montague
2012 Frontiers in Neurorobotics  
This contrasts with normative models of optimal control and reinforcement learning, which assume the existence of a value function that explains behavior, where (somewhat tautologically) behavior maximizes  ...  This implies that a sufficient explanation for valuable behavior is the accumulation of evidence for internal models of our world.  ...  We would like to thank our three reviewers for very helpful comments and guidance.  ... 
doi:10.3389/fnbot.2012.00011 pmid:23133414 pmcid:PMC3487150 fatcat:cmzbcikzszgnjdcyb3dyeo6bam

Perceptual decisions result from the continuous accumulation of memory and sensory evidence [article]

Aaron M. Bornstein, Mariam Aly, Samuel Feng, Nicholas B. Turk-Browne, Kenneth A. Norman, Jonathan D. Cohen
2017 bioRxiv   pre-print
Together, these results suggest that perceptual decisions arise from the continuous and integrated accumulation of memory and sensory evidence.  ...  Participants performed a cue-guided perceptual decision task with independently-varying memory and sensory evidence.  ...  This publication was made possible through the support of funding from the Intel corporation, and a grant from the John Templeton Foundation (Grant ID #57876; K.A.N. and J.D.C.).  ... 
doi:10.1101/186817 fatcat:tgpvglnqnbhxtaw6olk2btqota

Stochastic accumulation of feature information in perception and memory

Christopher Kent, Duncan Guest, James S. Adelman, Koen Lamberts
2014 Frontiers in Psychology  
or twice at study) and to new data.  ...  Many experiments on early stimulus perception involved (and continue to involve) presenting a stimulus for a brief period and then measuring the time to respond (RT).  ...  Conflict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest  ... 
doi:10.3389/fpsyg.2014.00412 pmid:24860530 pmcid:PMC4026707 fatcat:2hxsnq2vtjdhfngbdh4z5wfbpa

True eddy accumulation trace gas flux measurements: proof of concept

Lukas Siebicke, Anas Emad
2019 Atmospheric Measurement Techniques  
approaches to coordinate rotation, and uncertainties of vertical wind velocity measurements.  ...  </p> <p>Recently, a new dynamic and accurate mass flow controller which can resolve turbulence at a frequency of 10&amp;thinsp;Hz and higher has been developed by the first author.  ...  Lukas Siebicke would like to thank Thomas Foken for exposure to the theory of eddy accumulation flux measurements as well as Benoit Burban, Jean-Yves Goret, James Morison, Mike Perks, and Dorothea Schreiber  ... 
doi:10.5194/amt-12-4393-2019 fatcat:56d2lslltbfgzeoreva2udakne

Sparse Meta Networks for Sequential Adaptation and its Application to Adaptive Language Modelling [article]

Tsendsuren Munkhdalai
2020 arXiv   pre-print
The fast-weights are generated sparsely at each time step and accumulated incrementally through time providing a useful inductive bias for online continual adaptation.  ...  We introduce Sparse Meta Networks -- a meta-learning approach to learn online sequential adaptation algorithms for deep neural networks, by using deep neural networks.  ...  In this work, we explicitly train a deep neural network for such a fast and continual online adaptation. The core component of our approach is the Meta Networks with fast-weight memory.  ... 
arXiv:2009.01803v1 fatcat:52li37mdwvajlfmyc2hvuhcub4

Online parallel accumulation − serial fragmentation (PASEF) with a novel trapped ion mobility mass spectrometer [article]

Florian Meier, Andreas-David Brunner, Scarlet Koch, Heiner Koch, Markus Lubeck, Michael Krause, Niels Goedecke, Jens Decker, Thomas Kosinski, Melvin A Park, Nicolai Bache, Ole Hoerning (+3 others)
2018 bioRxiv   pre-print
MaxQuant identified more than 6,400 proteins in single run HeLa analyses without matching to a library, and with high quantitative reproducibility (R > 0.97).  ...  PASEF on the timsTOF Pro is a valuable addition to the technological toolbox in proteomics, with a number of unique operating modes that are only beginning to be explored.  ...  This work was partially supported by the German Research Foundation (DFG-Gottfried Wilhelm Leibniz Prize) granted to Matthias Mann and by the Max-Planck Society for the Advancement of Science.  ... 
doi:10.1101/336743 fatcat:j73fes72cfch3i6uzqkv7qvlti

On the nature of thought processes and their relationship to the accumulation of knowledge, Part XIII: The nature of evidence

Cris Anderson
2014 Dermatology Practical & Conceptual  
We looked at the work of Gerd Gigerenzer, who in Adaptive Thinking and Gut Feelings, explains the concept of "fast and frugal" heuristics and the role of Environment of Evolutionary Adaptation in the evolution  ...  True to pattern, when we find something, a method or an approach perhaps, we try it again and again until we finally are forced to admit that some sorts of problems require a different approach (the rules  ...  Learning to accept "evidence" for what it really represents instead of what we want it to represent will be a start.  ... 
doi:10.5826/dpc.0401a03 pmid:24520510 pmcid:PMC3919836 fatcat:4pja5pjafvbx3eeqshsbs5zlo4

Learning Fast and Slow for Online Time Series Forecasting [article]

Quang Pham, Chenghao Liu, Doyen Sahoo, Steven C.H. Hoi
2022 arXiv   pre-print
In this work, inspired by the Complementary Learning Systems (CLS) theory, we propose Fast and Slow learning Networks (FSNet), a holistic framework for online time-series forecasting to simultaneously  ...  Particularly, FSNet improves the slowly-learned backbone by dynamically balancing fast adaptation to recent changes and retrieving similar old knowledge.  ...  Fast and Slow Learning Networks (FSNet) By reconciling online time series forecasting with continual learning, we motivate from the CLS theory to develop a fast and slow learning framework for online time  ... 
arXiv:2202.11672v1 fatcat:ngkdygjcjfg5zfzky366idi3ei
« Previous Showing results 1 — 15 out of 45,349 results