Predicting Efficiency/Effectiveness Trade-offs for Dense vs. Sparse\n Retrieval Strategy Selection
Bibliographic record
Abstract
Over the last few years, contextualized pre-trained transformer models such\nas BERT have provided substantial improvements on information retrieval tasks.\nRecent approaches based on pre-trained transformer models such as BERT,\nfine-tune dense low-dimensional contextualized representations of queries and\ndocuments in embedding space. While these dense retrievers enjoy substantial\nretrieval effectiveness improvements compared to sparse retrievers, they are\ncomputationally intensive, requiring substantial GPU resources, and dense\nretrievers are known to be more expensive from both time and resource\nperspectives. In addition, sparse retrievers have been shown to retrieve\ncomplementary information with respect to dense retrievers, leading to\nproposals for hybrid retrievers. These hybrid retrievers leverage low-cost,\nexact-matching based sparse retrievers along with dense retrievers to bridge\nthe semantic gaps between query and documents. In this work, we address this\ntrade-off between the cost and utility of sparse vs dense retrievers by\nproposing a classifier to select a suitable retrieval strategy (i.e., sparse\nvs. dense vs. hybrid) for individual queries. Leveraging sparse retrievers for\nqueries which can be answered with sparse retrievers decreases the number of\ncalls to GPUs. Consequently, while utility is maintained, query latency\ndecreases. Although we use less computational resources and spend less time, we\nstill achieve improved performance. Our classifier can select between sparse\nand dense retrieval strategies based on the query alone. We conduct experiments\non the MS MARCO passage dataset demonstrating an improved range of\nefficiency/effectiveness trade-offs between purely sparse, purely dense or\nhybrid retrieval strategies, allowing an appropriate strategy to be selected\nbased on a target latency and resource budget.\n
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".