Zobrazeno 1 - 7
of 7
pro vyhledávání: '"Munjal, Prateek"'
Beyond Metrics: A Critical Analysis of the Variability in Large Language Model Evaluation Frameworks
Autor:
Pimentel, Marco AF, Christophe, Clément, Raha, Tathagata, Munjal, Prateek, Kanithi, Praveen K, Khan, Shadab
As large language models (LLMs) continue to evolve, the need for robust and standardized evaluation benchmarks becomes paramount. Evaluating the performance of these models is a complex challenge that requires careful consideration of various linguis
Externí odkaz:
http://arxiv.org/abs/2407.21072
Autor:
Christophe, Clément, Kanithi, Praveen K, Munjal, Prateek, Raha, Tathagata, Hayat, Nasir, Rajan, Ronnie, Al-Mahrooqi, Ahmed, Gupta, Avani, Salman, Muhammad Umar, Gosal, Gurpreet, Kanakiya, Bhargav, Chen, Charles, Vassilieva, Natalia, Amor, Boulbaba Ben, Pimentel, Marco AF, Khan, Shadab
This study presents a comprehensive analysis and comparison of two predominant fine-tuning methodologies - full-parameter fine-tuning and parameter-efficient tuning - within the context of medical Large Language Models (LLMs). We developed and refine
Externí odkaz:
http://arxiv.org/abs/2404.14779
Semantic segmentation from user inputs has been actively studied to facilitate interactive segmentation for data annotation and other applications. Recent studies have shown that extreme points can be effectively used to encode user inputs. A heat ma
Externí odkaz:
http://arxiv.org/abs/2004.02038
Active learning (AL) is a promising ML paradigm that has the potential to parse through large unlabeled data and help reduce annotation cost in domains where labeling data can be prohibitive. Recently proposed neural network based AL methods use diff
Externí odkaz:
http://arxiv.org/abs/2002.09564
Recently generative models have focused on combining the advantages of variational autoencoders (VAE) and generative adversarial networks (GAN) for good reconstruction and generative abilities. In this work we introduce a novel hybrid architecture, I
Externí odkaz:
http://arxiv.org/abs/1909.13062
Zero shot learning (ZSL) aims to recognize unseen classes by exploiting semantic relationships between seen and unseen classes. Two major problems faced by ZSL algorithms are the hubness problem and the bias towards the seen classes. Existing ZSL met
Externí odkaz:
http://arxiv.org/abs/1904.07659
Publikováno v:
2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).
Active learning (AL) is a promising ML paradigm that has the potential to parse through large unlabeled data and help reduce annotation cost in domains where labeling data can be prohibitive. Recently proposed neural network based AL methods use diff