The Effect of Information-Spread on Face Discrimination
Bibliographic record
Abstract
Despite our extensive experience with faces, we are surprisingly inefficient at face identification. Previous research in our lab and others has suggested that we use only a small proportion of the available information in face identification tasks, and that this information is centered about the eyes and eyebrows. Interestingly, the eye and brow regions are the most informative for face identification in our stimuli. Here we consider the possibility that observers are simply unable to process information across the entire face, and focus on localized regions as the best way to cope with this limitation. Observers discriminated between two faces in each of two different conditions. In one condition, all of the pixels were presented in localized regions around the eyes and brows (“high information value”). In the other, the pixels were distributed broadly about the face but did not include the same eye/brow regions (“low information value”). A staircase varied the total amount of information available in each condition by varying the number of pixels presented. For example, 10% of the stimulus information is packed into a relatively small number of pixels around the eyes/brows in the “high information value” condition, whereas 10% of the stimulus information in the “low information value” condition is spread about a much larger number of pixels. Observers required a significantly higher percent of information in the “low” condition than in the “high”, suggesting that the total amount of stimulus information is not as important as the spatial distribution of that information. Observers are much more efficient at discriminating faces based on the most informative regions, even when that information is contained in relatively few pixels. We are currently examining the effects of learning and stimulus context to determine the extent of flexibility in observers' face processing strategies.
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 | 0.002 | 0.026 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.006 | 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".