Dartmouth College Computer Science
Technical Report series
|By author:||A B C D E F G H I J K L M N O P Q R S T U V W X Y Z|
|By number:||2020, 2019, 2018, 2017, 2016, 2015, 2014, 2013, 2012, 2011, 2010, 2009, 2008, 2007, 2006, 2005, 2004, 2003, 2002, 2001, 2000, 1999, 1998, 1997, 1996, 1995, 1994, 1993, 1992, 1991, 1990, 1989, 1988, 1987, 1986|
We propose an algorithm for training neural networks in noisy label scenarios that up-weighs per-example gradients that are more similar to other gradients in the same minibatch. Our approach makes no assumptions about the amount or type of label noise, does not use a held-out validation set of clean examples, makes relatively few computations, and only modifies the minibatch gradient aggregation module in a typical neural network training workflow. For CIFAR-10 classification with varying levels of label noise, our method successfully up-weighs clean examples and de-prioritizes noisy examples, showing consistent improvement over a vanilla training baseline. Our results open the door to potential future work involving per-example gradient comparisons.
Senior Honors Thesis. Advisor: Lorenzo Torresani.
Bibliographic citation for this report: [plain text] [BIB] [BibTeX] [Refer]
Or copy and paste:
Jason Wei, "Label Noise Reduction Without Assumptions." Dartmouth Computer Science Technical Report TR2020-899, June 2020.
Notify me about new tech reports.
Search the technical reports.
To receive paper copy of a report, by mail, send your address and the TR number to reports AT cs.dartmouth.edu
Copyright notice: The documents contained in this server are included by the contributing authors as a means to ensure timely dissemination of scholarly and technical work on a non-commercial basis. Copyright and all rights therein are maintained by the authors or by other copyright holders, notwithstanding that they have offered their works here electronically. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.
Technical reports collection maintained by David Kotz.