Long-Context Efficient Transformers: A Comprehensive Survey of Techniques, Applications, and Future Directions
Bibliographic record
Abstract
Recent advancements in deep learning have led to the development of transformer models, which have achieved remarkable success in a wide variety of natural language processing (NLP) tasks. However, one of the major limitations of traditional transformers is their quadratic time and space complexity with respect to the input sequence length, making it difficult to process long-context information efficiently. This issue becomes especially pronounced when working with large documents, extended dialogues, or multimodal data. To address this challenge, longcontext efficient transformers have been introduced, which utilize innovative attention mechanisms to reduce computational complexity while retaining performance on long-context tasks. In this survey, we provide a comprehensive overview of the methods and models that have emerged in the field of long-context efficient transformers, including sparse attention mechanisms, kernel-based approaches, and memory-augmented models. We discuss the strengths and weaknesses of each approach and examine their real-world applications, including document summarization, multihop question answering, retrieval, and multimodal tasks. Additionally, we highlight the key challenges that remain in this area, such as the trade-offs between efficiency and expressiveness, generalization to longer sequences, and deployment considerations. Finally, we outline promising directions for future research aimed at overcoming these challenges and advancing the state of the art in long-context processing.
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 | 0.002 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.003 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.002 | 0.007 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.003 |
| Insufficient payload (model declined to judge) | 0.004 | 0.002 |
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".