LLM-Based Edge Intelligence: A Comprehensive Survey on Architectures, Applications, Security and Trustworthiness
Bibliographic record
Abstract
The integration of Large Language Models (LLMs) and Edge Intelligence (EI) introduces a groundbreaking paradigm for intelligent edge devices. With their capacity for human-like language processing and generation, LLMs empower edge computing with a powerful set of tools, paving the way for a new era of decentralized intelligence. Yet, a notable research gap exists in obtaining a thorough comprehension of LLM-based EI architectures, which should incorporate crucial elements such as security, optimization, and responsible development. This survey aims to bridge this gap by providing a comprehensive resource for both researchers and practitioners. We explore LLM-based EI architectures in-depth, carefully analyzing state-of-the-art paradigms and design decisions. To facilitate efficient and scalable edge deployments, we perform a comparative analysis of recent optimization and autonomy techniques specifically designed for resource-constrained edge environments. Additionally, we shed light on the extensive potential of LLM-based EI by demonstrating its varied practical applications across a wide range of domains. Acknowledging the utmost importance of security, our survey thoroughly investigates potential vulnerabilities inherent in LLM-based EI deployments. We explore corresponding defense mechanisms to protect the integrity and confidentiality of data processed at the edge. In conclusion, highlighting the essential aspect of trustworthiness, we outline best practices and guiding principles for the responsible development and deployment of these systems. By conducting a comprehensive review of these key components, our survey aims to support the ethical development and strategic implementation of LLM-driven EI, paving the way for its transformative impact on diverse applications.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.013 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.004 | 0.005 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.005 | 0.011 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".