Bibliographic record
Abstract
Abstract In recent years, attention transformers have proven to be instrumental in natural language processing (NLP)-based tasks such as sentence classification and language translation. However, their application has been recently extended to large-scale object recognition tasks. In this work, Vision Transformer with attention has been investigated for the detection of human falls and ADLs (Activities of Daily Living) from time series-based signals. The Vision Transformer model has been trained and validated using the acceleration signals of waist-worn Inertial Measurement Unit (IMU) sensors obtained from the IMU Falls dataset[1]. The model is also trained and validated on the popular SiSFall dataset[2]. The model is also investigated by independently training three different cases of patch size and attention heads. It was observed that a larger patch size resulted in significant performance deterioration. Additionally, a smaller patch size took longer to train and was computationally expensive. The model performed (best case) with an Accuracy (%) of 99.9 ± 0.1 and a True Positive Rate (%) of 99.9 ± 0.1 on the SFU-IMU dataset and with an Accuracy (%) of 99.8 ± 0.25 and a true positive rate (%) of 99.87 ± 0.3 on the SISFALL dataset. Overall, the results show that Transformers are highly robust in the detection of human falls and nonfalls/ADLs, subject to the appropriate patch size.
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 | 0.002 |
| 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 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 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".