A Language Model-based Approach to Context Analysis in Business English Translation
Bibliographic record
Abstract
In this paper, we use a large language model for business English translation and context analysis, and propose an adaptive parameter unfreezing method based on the quantization difference between adjacent layers within the decoder to fine-tune the layers of the language model related to the translation task, and to understand the behavior of the model in the relevant layers.Then the method of combining different encoders is proposed as a dual encoding-decoding framework on top of the traditional encoding-decoding framework, which is applied to the task of context analysis in business English translation.The fine-tuning method in this paper significantly improves the text translation quality of the language model, especially in the English-X tri-lingualization, which improves the COMET and BLEU metrics by 3.22 and 2.58 points respectively.In addition, the dual encodingdecoding model proposed in this paper is applicable to the task of contextual analysis in business English translation, which significantly improves the performance of contextual analysis in business English, and the F1 value on the HIT-CDTB dataset is improved by 11.60% compared with that of Rutherford's model.The experiment proves that the proposed method of text has made progress in the research of the task of analyzing textual contextual relations in business English.
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 | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 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".