Machine Learning for Pile-Up Decomposition in High Count Rate Gamma Spectroscopy Using a HPGe Detector
Bibliographic record
Abstract
In high count rate gamma-ray spectral measurements using a HPGe detector, pulse pile-up events cause a hard challenge for pulse height analysis due to distortions in pulse waveform and height. The conventional digital pulse height analyzers avoid the problems caused by the distortions by applying a pile-up rejection algorithm, which leads to unavoidable high dead time at high count rates. As a solution to address this issue, we present a convolutional neural network (CNN)-based method to decompose pile-up pulses from the preamplifier output and analyze the height of each decomposed pulse. The network is trained using the pile-up events that were artificially generated by applying various time intervals using clean pulses collected at low count rates using a planar HPGe detector. An input of the CNN model is the waveform of a pile-up pulse while the output comprises two pulses after decomposition, each of which is represented by a length of 2000 sampling channels per pulse at 16-bit precision. The height of each decomposed pulse is subsequently determined using a trapezoidal filter written in Python and a pulse height spectrum is produced. The pulse height spectrum obtained with the pile-up decomposition for the artificially generated pile-up pulse dataset demonstrated energy resolution comparable to the resolution of the spectrum collected using a commercial multi-channel analyzer. To evaluate the performance of the decomposition algorithm for real pile-up pulses from the HPGe detector, waveforms of pile-up pulses were collected at various count rates from 500 cps to 128 kcps and analyzed off-line through the CNN model for each count rate. The result showed a notable degradation in energy resolution at high count rates, primarily due to the baseline distortions caused by the incomplete decay of previous pulses. Further work in decomposition algorithm is currently underway to address this issue and improve the energy resolution.
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) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| 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) | 0.001 | 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".