Learning induced illusions: Statistical regularities create false memories
Bibliographic record
Abstract
Although the visual system readily extracts regularities in terms of object co-occurrences over space and time, does learning such statistical relationships always result in the veridical representations of individual objects? Here we investigate an interesting consequence of statistical learning: how does the knowledge of statistical regularities alter the representations of individual objects which no longer co-occur with each other? During the exposure phase, observers viewed a continuous sequence of objects while performing a cover task to ensure incidental encoding of the regularities. Unbeknownst to the observers, the objects appeared either in pairs (e.g., A always appeared before B) in the structured condition, or in a random order in the random condition. In a subsequent recognition phase, a new continuous sequence of objects was presented, and observers judged whether a specific object was present in the sequence. Importantly, the sequence now only contained one member of the original pair (e.g., only A was presented and B was missing), and observers judged whether A or B was present in the sequence. We found that observers in the structured condition showed a reliably higher false alarm rate for the missing object (e.g., B) than in the random condition. At the same time, the hit rate for the presented object (e.g., A) in the structured condition was also higher than in the random condition. The results demonstrate that statistical learning not only sharpens the detection of the object within the regularities, but also induces a false memory of the missing object. This finding reveals a novel consequence of statistical learning: learning that two objects co-occur can create the illusion of seeing one object, even though only its partner is present. Meeting abstract presented at VSS 2017
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 | 0.019 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| 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".