A Novel Optimized Deep Learning Model for Canola Crop Yield Prediction on Edge Devices
Bibliographic record
Abstract
The escalating global demand for food, coupled with challenges in sustaining crop production, deteriorating ocean health, and depleting natural resources, underscores the critical role of agricultural technology. This article addresses the imperative of developing an optimal deep-learning model for predicting canola crop yield using hyperspectral images captured by drone flights. Our primary objective is to identify the most efficient model in terms of performance and size, considering the storage limitations on edge devices like Raspberry Pi 4 (RPi4). We start with the baseline 1D<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">_</i>CNN model, which achieves an <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$R^{2}$</tex-math></inline-formula> score of 0.82, and compress it into the proposed <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">fs_model</i> (fp32). To achieve the compression, we apply pruning through sparsity and feature selection using SHAP values. Further reduction in model size is accomplished by quantizing the weights of the proposed model to a lower precision, such as int16. This combined approach substantially decreases the proposed model's size by approximately 92.6% and inference time by approximately ×9013 in comparison to the baseline 1D<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">_</i>CNN model. In addition, we propose the novel <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">fsp_model</i> posit(8,3) that uses posit quantization to further reduce the computation requirements compared to the proposed <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">fs_model</i> (int16). Our findings indicate that the utilization of posit numbers enables us to shrink the model size to 94% of the original base model, while only reducing the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$R^{2}$</tex-math></inline-formula> score by 5.7%.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 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 teacher head, 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".