Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

LLMs with Industrial Lens: Deciphering the Challenges and Prospects -- A Survey

Ashok Urlana, Charaka Vinayak Kumar, Ajeet Kumar Singh, Bala Mallikarjunarao Garlapati, Srinivasa Rao Chalamala, Rahul Mishra

TL;DR

This survey provides an industrial lens on large language models by combining an industry practitioner case study with a comprehensive review of 68 industry-focused papers. It identifies practical adoption drivers, datasets, evaluation metrics, and deployment challenges, and it documents a broad spectrum of applications from standard NLP tasks to security and governance. The work highlights gaps such as limited human evaluation, privacy concerns, and the need for multilingual and multimodal capabilities, offering a strategic roadmap to maximize industrial impact. By detailing datasets, prompting strategies, and evaluation approaches, the paper delivers actionable guidance for practitioners and researchers aiming to deploy LLMs responsibly in real-world settings.

Abstract

Large language models (LLMs) have become the secret ingredient driving numerous industrial applications, showcasing their remarkable versatility across a diverse spectrum of tasks. From natural language processing and sentiment analysis to content generation and personalized recommendations, their unparalleled adaptability has facilitated widespread adoption across industries. This transformative shift driven by LLMs underscores the need to explore the underlying associated challenges and avenues for enhancement in their utilization. In this paper, our objective is to unravel and evaluate the obstacles and opportunities inherent in leveraging LLMs within an industrial context. To this end, we conduct a survey involving a group of industry practitioners, develop four research questions derived from the insights gathered, and examine 68 industry papers to address these questions and derive meaningful conclusions. We maintain the Github repository with the most recent papers in the field.

LLMs with Industrial Lens: Deciphering the Challenges and Prospects -- A Survey

TL;DR

This survey provides an industrial lens on large language models by combining an industry practitioner case study with a comprehensive review of 68 industry-focused papers. It identifies practical adoption drivers, datasets, evaluation metrics, and deployment challenges, and it documents a broad spectrum of applications from standard NLP tasks to security and governance. The work highlights gaps such as limited human evaluation, privacy concerns, and the need for multilingual and multimodal capabilities, offering a strategic roadmap to maximize industrial impact. By detailing datasets, prompting strategies, and evaluation approaches, the paper delivers actionable guidance for practitioners and researchers aiming to deploy LLMs responsibly in real-world settings.

Abstract

Large language models (LLMs) have become the secret ingredient driving numerous industrial applications, showcasing their remarkable versatility across a diverse spectrum of tasks. From natural language processing and sentiment analysis to content generation and personalized recommendations, their unparalleled adaptability has facilitated widespread adoption across industries. This transformative shift driven by LLMs underscores the need to explore the underlying associated challenges and avenues for enhancement in their utilization. In this paper, our objective is to unravel and evaluate the obstacles and opportunities inherent in leveraging LLMs within an industrial context. To this end, we conduct a survey involving a group of industry practitioners, develop four research questions derived from the insights gathered, and examine 68 industry papers to address these questions and derive meaningful conclusions. We maintain the Github repository with the most recent papers in the field.
Paper Structure (39 sections, 7 figures, 6 tables)

This paper contains 39 sections, 7 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Related Surveys
  3. Benchmark Datasets
  4. Standard Datasets
  5. Application-specific Datasets
  6. Data generation using LLMs
  7. LLMs for Real-world Applications
  8. Standard NLP tasks
  9. Code Generation
  10. Trustworthy AI
  11. Retrieval & Recommendation
  12. Security
  13. Tools and Frameworks
  14. Societal Impact
  15. Evaluation Approaches
  16. ...and 24 more sections

Figures (7)

  • Figure 1: Number of papers from various applications, *entries merged with other application categories.
  • Figure 2: Structure of the paper.
  • Figure 3: Data modalities distribution for industrial applications.
  • Figure 4: Risks associated with LLMs for industrial applications.
  • Figure 5: Data modalities distribution for industrial applications; others includes privacy, bias, fairness and toxicity.
  • ...and 2 more figures