Overview
In academic research, finding literature requires balancing precision and coverage. Search systems that rely solely on keyword matching offer binary results: an article either contains the terms or it does not. In semantic search, however, relevance exists along a continuous spectrum of similarity in meaning space. A query on perovskite solar cells has no objective boundary where relevant publications end and irrelevant ones begin.
To give researchers direct control over this boundary, Keenious provides Scope. Scope acts as a semantic relevance floor, governing how closely matching a publication must be to enter the result set. Rather than forcing you into a single rigid retrieval cutoff, Scope lets you dial relevance from a tightly focused core to an open, exploratory survey of adjacent literature.
How Scope Works
Every publication title and abstract in the Keenious curated index of ~188 million records is represented as a high-dimensional vector (embedding) capturing its conceptual meaning. When you enter a search query, Keenious projects the query into this same meaning space and calculates the semantic distance between the query and potential literature matches.
Concentric relevance boundaries in meaning space around the search query, with the corresponding 4-segment meter.
Scope establishes concentric distance thresholds around your query:
Candidate Retrieval: When you run a query, Keenious retrieves candidate publications up to your chosen pool ceiling (300 by default, adjustable from 100 to 1,000).
Relevance Floor Evaluation: Keenious compares each candidate's semantic similarity score against the active Scope threshold.
Filtering: Publications meeting or exceeding the relevance floor are included in the result set. Publications falling below the floor are excluded.
Downstream Processing: All subsequent operations (research area clustering, citation sorting, and entity match counts) operate exclusively on the qualifying publications that pass this floor.
The Four Scope Levels
Scope offers four levels, represented in the interface by a segmented meter:
Level | Meter | Relevance Floor | Description | Recommended Usage |
Focused | 1 segment | High semantic similarity | Strict relevance floor. Only surfaces publications that directly address the core topic and methodology. | Deep dives into niche questions, verifying priority, or isolating specialized protocols. |
Balanced | 2 segments | Calibrated similarity threshold | Standard relevance floor (default). Balances topical precision with sufficient contextual breadth. | Initial literature reviews, broad topic overviews, and general research queries. |
Exploratory | 3 segments | Relaxed similarity threshold | Low relevance floor. Surfaces adjacent concepts, interdisciplinary connections, and peripheral applications. | Brainstorming cross-disciplinary methods, finding analogous problems, and exploring novel angles. |
All | 4 segments | No similarity floor | Bypasses the relevance floor entirely. Returns candidates ranked by rank fusion regardless of semantic distance. | Broadest possible discovery when a topic has limited literature or unusual phrasing. |
Reading the Scope Meter
The Scope control appears on the results bar beside the sort options and result count. It displays a four-segment meter alongside the active tier name (Focused, Balanced, Exploratory, or All). Each additional filled segment indicates an expanded boundary taking in a wider conceptual perimeter.
When you open the Scope dropdown menu, Keenious displays the exact number of matching publications available at each tier. These live candidate counts let you evaluate how many publications qualify before you switch tiers.
Scope and Candidate Pools
Keenious separates the candidate pool ceiling from the relevance floor:
Pool Ceiling ("Show up to"): Sets the maximum number of candidates Keenious evaluates for your search (100, 300, 500, or 1,000 for text queries).
Scope: Sets the minimum semantic similarity a publication must demonstrate to pass into the final results.
Because Scope eliminates candidates that fall below the relevance floor, the results count displays "Up to 300" (or your chosen size). If 142 publications meet the Balanced relevance floor within the retrieved candidates, Keenious presents exactly 142 publications. The count represents a verified total of publications meeting that relevance threshold, rather than an arbitrary truncation.
When browsing collections using filters alone without search keywords, Scope is disabled. Without query text, semantic similarity cannot be calculated, and results are ranked directly by the server using your chosen sort order.
How Scope Safeguards Sorting
Sorting without a relevance floor undermines discovery in semantic search. If a search system ranks results purely by citation count across an unbounded candidate list, landmark papers with tens of thousands of citations from neighboring disciplines inevitably dominate the top positions. A search for a specific chemical catalyst could easily be overwhelmed by general organic chemistry textbooks or statistical software manuals that mention the field tangentially.
Scope safeguards sorting by enforcing the relevance floor first:
Keenious filters candidates against your active Scope tier.
The remaining publications are guaranteed to meet the topical similarity threshold.
The Sort control (such as Most cited or Newest) reorders only this qualifying group.
By applying Scope before sorting, you can sort by citation count with confidence that every highly cited paper returned directly addresses your research question.
Scope Across Entity Lanes
Keenious provides specialized search lanes for discovering Authors, Venues, and Institutions alongside publications. In each entity lane, Scope controls the underlying publications evidence pool:
Authors: An author's ranking and match count reflect the publications in the evidence pool that satisfy the active Scope threshold. Switching from Balanced to Focused highlights specialists whose output is dedicated to the core question. Switching to Exploratory highlights researchers working across multidisciplinary intersections.
Venues: Journal and conference rankings reflect publication density within the scoped evidence pool. Focused scope identifies dedicated niche journals, while Exploratory scope identifies generalist journals publishing related work.
Institutions: University and research institute rankings aggregate researchers and publications within the scoped candidate pool, identifying leading centers of excellence for that specific topic.
Research Strategies
1. The Funnel Strategy: Landscape to Precision
When entering an unfamiliar field:
Begin in Balanced (default) to understand the central themes and terminology.
Open Research Areas to view the main clusters of current work.
Once you identify the specific subfield of interest, refine your query with exact terminology and switch Scope to Focused to isolate the core literature.
2. The Cross-Disciplinary Discovery Strategy
When seeking innovative solutions to a research bottleneck:
Run your query in Exploratory Scope.
Check adjacent Research Areas that appear outside your home discipline.
Examine papers from neighboring fields that share underlying mathematical models, experimental techniques, or conceptual frameworks.
3. Foundational Literature Audit
When compiling a comprehensive bibliography or literature review:
Set Scope to Focused.
Switch Sort to Most cited.
The resulting list surfaces the foundational, highly cited papers specific to your inquiry, protected against intrusion from unrelated mega-cited works.
4. Tracking Recent Developments
When monitoring recent output:
Set Scope to Balanced or Exploratory.
Switch Sort to Newest.
Review recent preprints and journal articles published in the current year.
Diagnostic Guide
Situation | Underlying Cause | Recommended Adjustment |
Search returns fewer results than the chosen pool size | The active Scope floor excluded candidates that did not meet the similarity threshold. | Broaden Scope to Exploratory or All, or increase query specificity. |
A known relevant publication is missing | The publication's semantic similarity score fell just below the active floor, or its topic is discussed only in the full text. | Switch Scope from Focused to Balanced or Exploratory. Verify that key terms appear in the paper's title or abstract. |
Results feel too broad or contain loosely related papers | The Scope tier is set too broadly for a narrow technical inquiry. | Switch Scope to Focused, or add quoted terms ( |
Scope control is greyed out | No search text is present in the query box. | Enter search keywords or phrases. Scope requires text to calculate semantic similarity. |