Designing Optimizer for Deep Learning (Record no. 16588)

MARC details
000 -LEADER
fixed length control field 01648nam a22001697a 4500
082 ## - DEWEY DECIMAL CLASSIFICATION NUMBER
Classification number 006.31
Item number M476D
100 ## - MAIN ENTRY--AUTHOR NAME
Personal name Mazumder, Sudipta
245 ## - TITLE STATEMENT
Title Designing Optimizer for Deep Learning
Statement of responsibility, etc by Sudipta Mazumder
260 ## - PUBLICATION, DISTRIBUTION, ETC. (IMPRINT)
Place of publication IIT Jodhpur
Name of publisher Department of Computer Science and Technology
Year of publication 2023
300 ## - PHYSICAL DESCRIPTION
Number of Pages vii,10p.
Other physical details HB
500 ## - GENERAL NOTE
General note Optimizers form the backbone of any convolutional network as they are responsible for making the functions converge faster. The optimizers do this by modifying the weights and learning rate of the algorithm, which reduces the loss and improves the accuracy. A lot of optimizers have gained traction over the years, out of which SGD and Adam take the cake. Adam has taken the lead as it helps to reduce the dying gradient problem of SGD. However, we still have scope for improvement. With this paper, we aim to introduce a new algorithm that surpasses the performance of Adam by calculating the angular gradients (cosine and tangent angles) at consecutive steps. This algorithm uses the gradient of the current step, the previous step, and the step previous to that. As we present more information to the optimizer’s algorithm, the algorithm has more information, making it better poised to make more accurate predictions at faster convergence rates. We have tested this approach on benchmark datasets and compared it with other state-of-the-art optimizers, and have obtained superior results in almost every approach.<br/>
650 ## - SUBJECT ADDED ENTRY--TOPICAL TERM
Topical Term Department of Computer Science and Technology
Topical Term Optimizers
Topical Term MTech Theses
700 ## - ADDED ENTRY--PERSONAL NAME
Personal name Paul, Angshuman
942 ## - ADDED ENTRY ELEMENTS (KOHA)
Koha item type Thesis
Holdings
Withdrawn status Lost status Damaged status Not for loan Collection code Permanent Location Current Location Shelving location Date acquired Source of acquisition Full call number Accession Number Price effective from Koha item type
        Theses S. R. Ranganathan Learning Hub S. R. Ranganathan Learning Hub Reference 2024-04-01 Office of Academics 006.31 M476D TM00524 2024-07-01 Thesis