Visual spread reading: Noisy letters in their natural context
Bibliographic record
Abstract
Keywords: Reading; Letters; Noise; Crowding. We quantified text legibility as a function of various factors present in natural texts. We adapted the visual spread method (Poirier, Gosselin & Arguin, submitted) to a reading task. Stimuli were sentences conforming to MN Read standards (Legge, Ross & Luebker, 1989, O&VS) mixed with dynamic probabilistic noise - i.e. each pixel in the image was associated with a probability that its polarity was inverted on a given refresh cycle of the display screen. Even low levels of noise can alter letter shapes. Noise level varied continuously over the image as initially determined by Gaussian-filtered noise. Participants adjusted noise levels in the text using the mouse until the text appeared homogenously noisy. We assumed that participants increased (or decreased) noise at locations where stimulus features were easy (or difficult) to encode, thus that local noise setting would correlate with legibility. Data on 5 participants and 10 texts revealed interesting effects; demonstrating the method's validity for assessing legibility in natural texts. For example, participants put more noise (1) over spaces than over letters, (2) over the first and last letter of words than over middle letters, and (3) over letters that are highly discriminable (e.g. for the 15 letters that appear at least 10x in the texts, r = .45–.61 with confusability data using traditional psychophysical techniques; see Fiset, Dupuis-Roy, Arguin & Gosselin, in preparation). Critically, these preliminary analyses demonstrate the sensitivity of the method to the legibility of individual letters within its natural context. In-depth analyses will include various other factors, e.g. letter and word frequency, word length and type, letter position in the word, as well as possible interactions between these factors.
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 | 0.015 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 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 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".