2,953 Hits in 10.7 sec

Where Did You Learn That From? Surprising Effectiveness of Membership Inference Attacks Against Temporally Correlated Data in Deep Reinforcement Learning [article]

Maziar Gomrokchi, Susan Amin, Hossein Aboutalebi, Alexander Wong, Doina Precup
2021 arXiv   pre-print
Furthermore, we study the differences in the performance of collective and individual membership attacks against deep reinforcement learning algorithms.  ...  Experimental results show that the proposed adversarial attack framework is surprisingly effective at inferring the data used during deep reinforcement training with an accuracy exceeding 84% in individual  ...  ACKNOWLEDGEMENTS The authors would like to thank Hamidreza Ghafghazi for his valuable contribution to the design and development of the preliminary version of the codebase.  ... 
arXiv:2109.03975v2 fatcat:aktgp5aefzcj5hs7acymclg23m

Investigating Membership Inference Attacks under Data Dependencies [article]

Thomas Humphries, Simon Oya, Lindsey Tulloch, Matthew Rafuse, Ian Goldberg, Urs Hengartner, Florian Kerschbaum
2021 arXiv   pre-print
One such attack, the Membership Inference Attack (MIA), exposes whether or not a particular data point was used to train a model.  ...  Training machine learning models on privacy-sensitive data has become a popular practice, driving innovation in ever-expanding fields.  ...  Where Did You Learn That From?  ... 
arXiv:2010.12112v3 fatcat:xrv65rf4yrdzxjvsoy5y22irm4

Report from the NSF Future Directions Workshop, Toward User-Oriented Agents: Research Directions and Challenges [article]

Maxine Eskenazi, Tiancheng Zhao
2020 arXiv   pre-print
In the breakout sessions that followed, the participants defined the main research areas within the domain of intelligent agents and they discussed the major future directions that the research in each  ...  Any opinions, findings and conclusions or future directions expressed in this document are those of the authors and do not necessarily reflect the views of the National Science Foundation.  ...  Learning from limited data (column 2) Machine learning methods, especially supervised deep learning models require a tremendous amount of training data to reach good performance.  ... 
arXiv:2006.06026v1 fatcat:h2hbe4cr6vajvca4oqkfrmf5oa

Finding epic moments in live content through deep learning on collective decisions

Hyeonho Song, Kunwoo Park, Meeyoung Cha
2021 EPJ Data Science  
The evaluation shows that our data-driven approach can identify epic moments from user-generated streamed content that cover various contexts (e.g., victory, funny, awkward, embarrassing).  ...  We characterize what features "epic" moments and present a deep learning model to extract them based on analyzing two million user-recommended clips and the associated chat conversations.  ...  Figure 8 (c) is a scene where the broadcaster is suddenly attacked inside the game, but he successfully defended against the attack. The scene was described as "That was a surprising turnaround.  ... 
doi:10.1140/epjds/s13688-021-00295-6 fatcat:sjuoj2chpbalzhstzhicipbhfe

Democracy from above: Regional Organizations and Democratization

G. John Ikenberry, Jon C. Pevehouse
2005 Foreign Affairs  
In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted.  ...  Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.  ...  If, in fact, these are the only cases where this mechanism functions, then it is not surprising that there is no association between lOs and the completion o f democracy in the Gasiorowski data, which  ... 
doi:10.2307/20034358 fatcat:qhbdwfkh3jesdejwfvb5cp3qxi

With diversity in mind: Freeing the language sciences from Universal Grammar

Nicholas Evans, Stephen C. Levinson
2009 Behavioral and Brain Sciences  
In fact, there are vanishingly few universals of language in the direct sense that all languages exhibit them. Instead, diversity can be found at almost every level of linguistic organization.  ...  This fundamentally changes the object of enquiry from a cognitive science perspective.  ...  ., Chinese, Japanese, Lakhota) move their wh-forms, saying, in effect, "You came to see who?" instead of "Who did you come to see _" (Van Valin & LaPolla 1997, pp. 424 -25) .  ... 
doi:10.1017/s0140525x09990525 fatcat:pqieza6vancijg4vfnu7humrli

Advances and Open Problems in Federated Learning [article]

Peter Kairouz, H. Brendan McMahan, Brendan Avent, Aurélien Bellet, Mehdi Bennis, Arjun Nitin Bhagoji, Kallista Bonawitz, Zachary Charles, Graham Cormode, Rachel Cummings, Rafael G.L. D'Oliveira, Hubert Eichner (+47 others)
2021 arXiv   pre-print
FL embodies the principles of focused data collection and minimization, and can mitigate many of the systemic privacy risks and costs resulting from traditional, centralized machine learning and data science  ...  Federated learning (FL) is a machine learning setting where many clients (e.g. mobile devices or whole organizations) collaboratively train a model under the orchestration of a central server (e.g. service  ...  [183] recently showed that multiple Byzantine-resilient defenses did little to defend against model poisoning attacks in federated learning.  ... 
arXiv:1912.04977v3 fatcat:efkbqh4lwfacfeuxpe5pp7mk6a

Deep Learning in Information Security [article]

Stefan Thaler, Vlado Menkovski, Milan Petkovic
2018 arXiv   pre-print
Deep Learning is a sub-field of machine learning, which uses models that are composed of multiple layers.  ...  Consequently, representations that are used to solve a task are learned from the data instead of being manually designed.  ...  A uror: defending against poisoning attacks in collaborative deep learning systems.  ... 
arXiv:1809.04332v1 fatcat:xfb7lgrkw5cirdl3qvmg3ssnbi

11. Care Leavers in Italy [chapter]

Laura Formenti, Andrea Galimberti, Mirella Ferrari
2020 Discourses We Live By  
Laura Formenti, Andrea Galimberti and Mirella Ferrari use data from their study of young people leaving the residential care system on reaching adulthood.  ...  Each researcher demonstrates a distinctive approach – through realism, by focusing on a facet of interest, by fictionalizing the interview in order to go beyond it.  ...  Perhaps one question we will be asked is where the film is stored, and can you guarantee that it would be available to readers for the lifetime of the book?  ... 
doi:10.11647/obp.0203.11 fatcat:oy27n3y2zbhtjpyu64qffgxjy4

A review of machine learning applications in wildfire science and management [article]

Piyush Jain, Sean C P Coogan, Sriram Ganapathi Subramanian, Mark Crowley, Steve Taylor, Mike D Flannigan
2020 arXiv   pre-print
There exists opportunities to apply more current ML methods (e.g., deep learning and agent based learning) in wildfire science.  ...  Finally, we stress that the wildfire research and management community plays an active role in providing relevant, high quality data for use by practitioners of ML methods.  ...  Acknowledgments The motivation for this paper arose from the "Not the New Normal" BC AI Wildfire Symposium held in Vancouver, BC, on 12 October 2018.  ... 
arXiv:2003.00646v1 fatcat:5ufhtbwlsvd2rdk3ogbmqpnxuu

From Supervenience to Superdupervenience: Meeting the Demands of a Material World

1993 Mind  
Rene DePlanque of Fachinformationszentrum Chemie, Berlin, for providing us with the reaction data set used in evaluating the hierarchical classification algorithm.  ...  Colombetti for many valuable suggestions, and for long and stimulating discussions; also to Paolo Giangrandi, Elena Not, Fabio Rinaldi, and two anonymous referees, for many useful comments on an earlier draft of  ...  Consequently from its inception, his unit was a target of political attacks from other components which did not understand the fact that the faculty members of his unit worked several times harder than  ... 
doi:10.1093/mind/102.408.555 fatcat:aqjmty7ylvhxdeiuhihi2un7r4

Health Benefits from Energy Access in LMICs [chapter]

Nigel Bruce, Chen Ding
2014 Energy Poverty  
We have also described evidence that demonstrates the extent of the descriptive correlation of education and health.  ...  They find that an increase in the average number of years of education in the household reduces child mortality by approximately 10 percentage points from a mean level of 22.5%.  ...  Some of the benefits and/or risks of education for health result from the membership or participation in a learning institution, others from the explicit experience of the process of learning.  ... 
doi:10.1093/acprof:oso/9780199682362.003.0006 fatcat:kffcshj4sndf5lfglyak7vv5si

Papers read at the 15th Annual Meeting of the Psychonomic Society

1974 Bulletin of the Psychonomic Society  
Change in meaning (inferred from free associations) of more stimuli was correlated with easier learning only in the A-D paradigm. 2:40·2:55 (74) Individual Differences in Paired· Associate Learning.  ...  Effects are greater with passages containing a larger number of ditferent nouns. 11 :30-11 :40 (276) When You Know That You Know and What You Think That You Know But You Don't. EUGENE B. ZECHMEISTER.  ...  The findings indicate that the conditioning history is an important determinant in the manner in which humans mediate a temporally spaced (DRL) schedule of reinforcement SUNY.  ... 
doi:10.3758/bf03336721 fatcat:hliyw3nyg5dadi7koywcyl66ha

Social physics [article]

Marko Jusup, Petter Holme, Kiyoshi Kanazawa, Misako Takayasu, Ivan Romic, Zhen Wang, Suncana Gecek, Tomislav Lipic, Boris Podobnik, Lin Wang, Wei Luo, Tin Klanjscek (+3 others)
2021 arXiv   pre-print
Recent decades have seen a rise in the use of physics-inspired or physics-like methods in attempts to resolve diverse societal problems.  ...  Such a rise is driven both by physicists venturing outside of their traditional domain of interest, but also by scientists from other domains who wish to mimic the enormous success of physics throughout  ...  Inference algorithms. Algorithms that effectively infer community affiliations often come from the Monte-Carlo class of methods used in statistical physics.  ... 
arXiv:2110.01866v1 fatcat:ccfxyezl6zgddd6uvrxubmaxua

From Past to Present: The Deep History of Kinship [chapter]

Dwight W. Read
2019 Computational Social Sciences  
of case studies from numerous regions of the world and all prehistoric or historic periods.  ...  methodologies from the computational and complex systems sciences in order to promote deeper understanding and collaboration within the study of past human behavior and history.  ...  We thank our colleagues Ricardo del Olmo and David Poza, from Universidad de Burgos for their help in programming a preliminary version of our model.  ... 
doi:10.1007/978-3-030-12723-7_6 fatcat:mpv5ufnwtrgjzaq3xps5fk5e6u
« Previous Showing results 1 — 15 out of 2,953 results