The brain’s representations may be compatible with convolution-based memory models.
Bibliographic record
Abstract
Convolution is a mathematical operation used in vector-models of memory that have been successful in explaining a broad range of behaviour, including memory for associations between pairs of items, an important primitive of memory upon which a broad range of everyday memory behaviour depends. However, convolution models have trouble with naturalistic item representations, which are highly auto-correlated (as one finds, e.g., with photographs), and this has cast doubt on their neural plausibility. Consequently, modellers working with convolution have used item representations composed of randomly drawn values, but introducing so-called noise-like representation raises the question how those random-like values might relate to actual item properties. We propose that a compromise solution to this problem may already exist. It has also long been known that the brain tends to reduce auto-correlations in its inputs. For example, centre-surround cells in the retina approximate a Difference-of-Gaussians (DoG) transform. This enhances edges, but also turns natural images into images that are closer to being statistically like white noise. We show the DoG-transformed images, although not optimal compared to noise-like representations, survive the convolution model better than naturalistic images. This is a proof-of-principle that the pervasive tendency of the brain to reduce auto-correlations may result in representations of information that are already adequately compatible with convolution, supporting the neural plausibility of convolution-based association-memory. (PsycINFO Database Record
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 | 0.011 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.003 |
| Scholarly communication | 0.004 | 0.010 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.011 | 0.002 |
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".