Improving Large Language Model Performance Through Compression and Optimization
Bibliographic record
Abstract
The rapid advancement of large language models (LLMs) has driven remarkable improvements in natural language processing (NLP) applications. However, their escalating size and computational complexity pose significant challenges for real-world deployment, particularly in latency-sensitive, resource-constrained, and energy-limited environments. Model compression techniques have emerged as a vital solution to mitigate these challenges by reducing model size, enhancing inference speed, and improving memory efficiency without compromising performance. This survey provides a comprehensive overview of state-of-the-art compression strategies tailored for LLMs, including pruning, quantization, knowledge distillation, low-rank approximation, and weight sharing. We also explore hardware-aware optimizations that leverage specialized hardware accelerators such as GPUs, TPUs, and FPGAs to maximize efficiency. Furthermore, we examine inference frameworks and libraries that facilitate efficient model deployment across diverse environments. In addition to covering established techniques, this survey highlights emerging trends in parameter-efficient fine-tuning and adaptive compression strategies that dynamically adjust model complexity based on input characteristics. We also discuss the trade-offs associated with various compression methods, including accuracy loss, hardware compatibility, and deployment overhead. Finally, we outline key open challenges in model compression, such as improving generalization in compressed models, ensuring energy-efficient deployment, and enhancing privacy and security. By consolidating insights from recent advancements, this survey aims to provide researchers and practitioners with a comprehensive guide to designing efficient language models capable of delivering high performance in resource-constrained settings.
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 | 0.021 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.002 | 0.003 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.008 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.001 | 0.003 |
| Insufficient payload (model declined to judge) | 0.004 | 0.003 |
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".