Bibliographic record
Abstract
We present a proposal to answer: how could an agent learn to better attend to the relevant and ignore the irrelevant in the context of performing visual tasks? To answer this, attention is defined as in Tsotsos (2011): Attention is the set of mechanisms that tune and control the search processes inherent in perception and cognition, dynamically adapting a general purpose processor to the input and task of the moment. To improve one's attention means that tuning and search control become more effective, e.g., task performance shows improvements in speed and accuracy. The link between this definition of attention and such improvements lies in the computational foundations underlying the Selective Tuning (ST) attentional theory, namely computational complexity (Tsotsos et al. 1995; Tsotsos 1990, 2011). Suppose one compares two algorithms, both effective for the same problem. The one with lower time complexity will lead to a faster solution. Lower time requirements can be achieved by reducing the number of candidates to consider via task-driven suppression or grouping. Given the same amount of time, two algorithms can be compared in terms of their accuracy. The more accurate algorithm will have improved decision-making mechanisms, perhaps by reducing the impacts of noise, ambiguity, or number of potential choices of action, or by eliminating interfering computations. For example, stronger suppression of distractors may reduce the impact of noise. The set of attentive mechanisms (selection, suppression and restriction and their 15 sub-classes) within ST are each examined and their variations with respect to performance (time and accuracy) are built into an overall optimization criterion that drives any changes due to experience. A Hebbian learning strategy is used to combine the minimization of time complexity while maximizing accuracy in a neurobiologically plausible manner. Finally, we point to experimental work that might verify the proposal. Meeting abstract presented at VSS 2014
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 | 0.004 | 0.022 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.005 |
| Scholarly communication | 0.005 | 0.012 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.005 | 0.004 |
| Insufficient payload (model declined to judge) | 0.016 | 0.004 |
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".