Triple-Aware Reasoning: A Retrieval-Augmented GenerationApproach for Enhancing Question-Answering Tasks withKnowledge Graphs and Large Language Models
Bibliographic record
Abstract
With the rapid developments of large language models, the application of large language models has spread to almost every aspect of human life. However, large language models also face some challenges. Large language models may be limited by the data used during pre-training and cannot obtain the latest knowledge. In addition, due to the huge amount of training data during the pre-training, it is sometimes difficult for large language models to pay more attention to knowledge with more semantic information for specific domains. To address these issues, while utilizing the state-of-the-art large language models, we present a retrieval-augmented generation-based method called Triple-Aware Reasoning. This method involves searching the knowledge graph for all alternative answer triples and subjecting the triples to multiple filtering to ensure the relevance of external knowledge and selecting relevant answers for each question-answering pair. By integrating knowledge graphs into the reasoning process, our method enables large language models to access and utilize additional external knowledge. Experimental results on the well-known CommonsenseQA and OpenBookQA datasets demonstrate the effectiveness of our approach. The large language model combined with the retrieval-augmented generation method demonstrates higher accuracy than the original model in multiple-choice question-answering tasks. The experimental results show the improvement by our method on the inference performance of large language models.
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 | 0.006 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.003 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.005 |
| Open science | 0.004 | 0.003 |
| Research integrity | 0.002 | 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".