Rehan Guha

Mr. Rehan Guha

orcid

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
15 January 2016 in International Journal of Computer Applications Technology and Research
Verified
Journal Article
Closed access
Rehan GuhaRehan Guha
Rehan Guha
1

1 January 2023 in SSRN Electronic Journal
Verified
Preprint
Open access
PDF file
Rehan GuhaRehan Guha
Rehan Guha
In recent years after emerging of LLM in the space of NLP, tools like ChatGPT, BARD, DALL-E, took over the market and daily lives. We will be focusing on the human like output capability from LLM’s . There is a huge reservation of this technology due to multiple scenarios like security, accuracy, relevance, etc... In this paper, I will talk about a method which I have designed to fine tune a human being to lower the expectation from LLM outputs and increase the acceptance rate of the final product. This technique is more of a psychological method than a technological way to improve the models output to be more human like.

6 January 2023 in Preprints.org
Verified
Preprint
Open access
PDF file
Rehan GuhaRehan Guha
Rehan Guha
Initial choice of Learning Rate is a key part of gradient based methods and has a great effect on the performance of the Deep Learning Model. This paper studies the behavior of multiple gradient based optimization algorithm which are commonly used in Deep Learning and compare their performance on various learning rate. As observed popular choice of optimization algorithms are highly sensitive to various choice of learning rates. Our goal is to find which optimizer has an edge over others for a specific setting. We look at two datasets namely MNIST and CIFAR10 for benchmarking. The results are quite surprising, and it will help us to choose a learning rate more efficiently.

5 January 2023 in ScienceOpen Preprints
Verified
Preprint
Open access
PDF file
Rehan GuhaRehan Guha
Rehan Guha
Initial choice of Learning Rate is a key part of gradient based methods and has a great effect on the performance of the Deep Learning Model. This paper studies the behavior of multiple gradient based optimization algorithm which are commonly used in Deep Learning and compare their performance on various learning rate. As observed popular choice of optimization algorithms are highly sensitive to various choice of learning rates. Our goal is to find which optimizer has an edge over others for a specific setting. We look at two datasets namely MNIST and CIFAR10 for benchmarking. The results are quite surprising, and it will help us to choose a learning rate more efficiently.

5 January 2023 in OSF Preprints
Verified
Preprint
Open access
Rehan GuhaRehan Guha
Rehan Guha
Initial choice of Learning Rate is a key part of gradient based methods and has a great effect on the performance of the Deep Learning Model.This paper studies the behavior of multiple gradient based optimization algorithm which are commonly used in Deep Learning and compare their performance on various learning rate. As observed popular choice of optimization algorithms are highly sensitive to various choice of learning rates. Our goal is to find which optimizer has an edge over others for a specific setting. We look at two datasets namely MNIST and CIFAR10 for benchmarking. The results are quite surprising, and it will help us to choose a learning rate more efficiently.

20 September 2023 in OSF Preprints
Verified
Preprint
Open access
Rehan GuhaRehan Guha
Rehan Guha
In recent years after emerging of LLM in the space of NLP, tools like ChatGPT, BARD, DALL-E, took over the market and daily lives. We will be focusing on the human like output capability from LLM’s . There is a huge reservation of this technology due to multiple scenarios like security, accuracy, relevance, etc... In this paper, I will talk about a method which I have designed to fine tune a human being to lower the expectation from LLM outputs and increase the acceptance rate of the final product. This technique is more of a psychological method than a technological way to improve the models output to be more human like.