Filters








3,652 Hits in 1.2 sec

Graph Unlearning [article]

Min Chen and Zhikun Zhang and Tianhao Wang and Michael Backes and Mathias Humbert and Yang Zhang
2021 arXiv   pre-print
In this paper, we propose GraphEraser, a novel machine unlearning method tailored to graph data.  ...  However, directly applying SISA to the graph data can severely damage the graph structural information, and thereby the resulting ML model utility.  ...  Figure 4 illustrates the node unlearning efficiency for different graph unlearning methods.  ... 
arXiv:2103.14991v1 fatcat:dwx425ody5eireo6w6sz3mrzwm

Making Recommender Systems Forget: Learning and Unlearning for Erasable Recommendation [article]

Yuyuan Li, Xiaolin Zheng, Chaochao Chen, Junlin Liu
2022 arXiv   pre-print
As machine learning models potentially memorize the training data, data erasure should also unlearn the data lineage in models, which raises increasing interest in the problem of Machine Unlearning (MU  ...  Both theoretical analysis and experiments on two real-world datasets demonstrate that LASER can not only achieve efficient unlearning, but also outperform the state-of-the-art unlearning framework in terms  ...  A.1 Preliminary of Hypergraph Agarwal et al. [2006] showed that the hypergraph with edgeindependent vertex weight can be reduced to either clique graph or star graph.  ... 
arXiv:2203.11491v1 fatcat:2mlhnb2uubdkviesgyb3pbytra

Recommendation Unlearning [article]

Chong Chen, Fei Sun, Min Zhang, Bolin Ding
2022 arXiv   pre-print
Experimental results on three public benchmarks show that RecEraser can not only achieve efficient unlearning, but also outperform the state-of-the-art unlearning methods in terms of model utility.  ...  The source code can be found at https://github.com/chenchongthu/Recommendation-Unlearning  ...  [13] applied this idea to the unlearning of graph with an improved sharding algorithm.  ... 
arXiv:2201.06820v2 fatcat:hwx2krxsvjcrjiwenrafiuzxzm

Bayesian Variational Federated Learning and Unlearning in Decentralized Networks [article]

Jinu Gong, Osvaldo Simeone, Joonhyuk Kang
2021 arXiv   pre-print
The proposed protocols are demonstrated to yield efficient unlearning mechanisms.  ...  This paper studies federated learning and unlearning in a decentralized network within a Bayesian framework.  ...  For a complete graph, since the probability that we first reach to agent k at l-th (l ≥ 2) iteration is (K − 1)/K × ((K − 2)/(K − 1)) l−2 × 1/(K − 1) , the expected number of iterations for unlearning  ... 
arXiv:2104.03834v1 fatcat:tdve5tq7wzaavdmaurnqdasvte

Page 9626 of The Journal of Neuroscience Vol. 29, Issue 39 [page]

2008 The Journal of Neuroscience  
The inset graph shows mean data across all cells tested (n = 5).  ...  We refer to this active process as unlearning. Unlearning occurs in any standard TD model but alone cannot allow for spontaneous recovery (Montague et al., 1996).  ... 

Page 728 of The Living Age Vol. 173, Issue 2243 [page]

1887 The Living Age  
728 Therein they failed, judged by the mono- graph standard, but gave a subtle food for the mind.  ...  Then to unlearn the first ideas of his- tory, of science, of social institutions, to unlearn one’s own life and purpose; to unlearn the old mode of thought aiid way of arriving at things ; to take off  ... 

Variational Bayesian Unlearning [article]

Quoc Phong Nguyen, Bryan Kian Hsiang Low, Patrick Jaillet
2020 arXiv   pre-print
This paper studies the problem of approximately unlearning a Bayesian model from a small subset of the training data to be erased.  ...  the posterior belief given the full data (i.e., including the remaining data); the latter prevents catastrophic unlearning that can render the model useless.  ...  Graphs of averaged KL divergence vs. λ achieved by EUBO, rKL, and q(θ|D) (i.e., baseline labeled as full) over (d) D r and (e) D e .  ... 
arXiv:2010.12883v1 fatcat:2jfmbvhyk5eitlhpqnh5k7vlda

Page 131 of Behavior Research Methods Vol. 4, Issue 3 [page]

1972 Behavior Research Methods  
Percentage of total opportunities for each aerial position where the initial presentation of the prior item cue (top graph) and of the position cue (bottom graph) elicited the correct response. .  ...  Furthermore, in the present study the presentation of unlearned prior item as a probe did not appear to contribute significantly to SL, since such instances occurred only on the average of 1.6 times per  ... 

Transformative Learning: The Role of Research in Traditional Clinical Disciplines

Michael Tims
2014 Integrative Medicine  
MUIH's methods include engaging students in visual deconstruction of research data found in graphs, tables, and images; having students identify the end point of a line of inquiry; and evaluating a previously  ...  integrating research methods into class and group projects can accelerate the unlearning process.  ...  projects can accelerate the unlearning process.  ... 
pmid:26770105 pmcid:PMC4566448 fatcat:lur3opnhizhmrdoz4j4fpl3ofm

Page 15 of International Pharmaceutical Abstracts Vol. 8, Issue 3 [page]

1971 International Pharmaceutical Abstracts  
Numerous graphs illustrating effects are included. (20 references) TousiGNauTAA® 8 0448 INHIBITORY PROPERTIES OF AMINOALKYLSUCCINIMIDES ON ISOLATED HORSE SERUM BUTYRYLCHOLINESTERASE: N-METHYL-2- PHENYL  ...  The tranquilizer, chlorpromazine, required a debilitating dose to block 4 of the 5 unlearned behaviors.  ... 

Unlearning in health care

R Rushmer
2004 BMJ Quality & Safety  
phenomena N Latent learning: where we soak up capacity, but this only becomes apparent when used at a later time Learning curve 37 (how we learn through time) A line graph where performance is plotted  ...  unlearner drops.  ... 
doi:10.1136/qhc.13.suppl_2.ii10 pmid:15576685 pmcid:PMC1765805 fatcat:i7rcsvoynzhyjm25gaj2zwinpm

Unlearning in health care

R Rushmer
2004 BMJ Quality & Safety  
phenomena N Latent learning: where we soak up capacity, but this only becomes apparent when used at a later time Learning curve 37 (how we learn through time) A line graph where performance is plotted  ...  unlearner drops.  ... 
doi:10.1136/qshc.2003.009506 pmid:15576685 pmcid:PMC1765805 fatcat:i6iaprt4y5cvxbslovg2nklfbm

Zero-Shot Machine Unlearning [article]

Vikram S Chundawat, Ayush K Tarun, Murari Mandal, Mohan Kankanhalli
2022 arXiv   pre-print
We therefore ask the question: is it possible to achieve unlearning with zero training samples?  ...  We also introduce a new evaluation metric, Anamnesis Index (AIN) to effectively measure the quality of the unlearning method.  ...  The graphs show how stopping after the right number of epochs is important. (Tables1 and 2present the results just before the accuracy on D f shoots up.)  ... 
arXiv:2201.05629v1 fatcat:7vrtdxqpdvg67nd23t3chcbpya

Learn to Forget: Machine Unlearning via Neuron Masking [article]

Yang Liu, Zhuo Ma, Ximeng Liu, Jian Liu, Zhongyuan Jiang, Jianfeng Ma, Philip Yu, Kui Ren
2021 arXiv   pre-print
first uniform metric called for-getting rate to measure the effectiveness of a machine unlearning method.  ...  We also propose a novel unlearning method calledForsaken. It is superior to previous work in either utility or efficiency (when achieving the same forgetting rate).  ...  The labels of the points in the graph are based on the actual values, not membership oracle, and do not change with machine unlearning. of I.C. after machine unlearning.  ... 
arXiv:2003.10933v3 fatcat:spjwz2chi5ecld7wdskt2adtbu

Page 28 of Arena Vol. 13, Issue 1 [page]

1895 Arena  
hands and go even the length of nationalizing railroads, tele- graphs, etc.  ...  I will enumerate a few of the things to be learned or unlearned, and in doing so, follow rather the analytical than the synthetic method.  ... 
« Previous Showing results 1 — 15 out of 3,652 results