Thick Data Analytics for Identifying Eye Conditions using Siamese Lookalike Neural Networ
Bibliographic record
Abstract
Human’s perceptual ruling on image similarity uses a phenomenon-learning model called pareidolia to recognize shapes like faces from in random samples. This short paper is an attempt to present a learning model that has the capability to recognize eye conditions from small training sample. The proposed model uses a LookALike learning technique to attempt providing higher perceptual capability that goes beyond the normal supervised learning. The proposed model employs a Siamese neural net with triplet loss function as well as two thick data analytics methods involving retina vessels augmentation and dropout layers. The model is trained on 400 eye conditions dataset from Kaggle where it distributed on four categories (cataract, glaucoma, Diabetic Retinopathy and Retina_Disease) as well as generative model to distribute lookalike eye conditions as closely similar to anchor, positively related to anchor or negatively related to anchor). The LookALike learning model shows 60% accuracy tested on 100 new eye cases and is improved to 70% accuracy when we added the iris vessels augmentation and other dropout filters. We are intending to use more thick data learning techniques to improve its performance among other research investigations, which we leave it to our future research.
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 | 0.003 |
| 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 | 0.001 |
| 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 | 0.001 |
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".