heroxbd waveform-analysis issues

heroxbd / waveform-analysis

Waveform analysis for reviewing all the effective methods.

MIT License

9 stars 4 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Proof of publication round 2

#163 heroxbd closed 2 years ago
5
Proof of publication

#162 heroxbd closed 2 years ago
61
CuPy 没有对 fancy index 报 overflow

#161 Berrysoft closed 2 years ago
0
minor fix to reply

#160 littlewu2508 closed 2 years ago
0
Change to real FSMP.

#159 Berrysoft closed 2 years ago
0
suggestions to the reply

#158 littlewu2508 closed 2 years ago
0
Merge newer changes from master for xGPU

#157 Berrysoft closed 2 years ago
1
back-propagation 引用 Hilton 的论文

#156 heroxbd closed 2 years ago
1
Collection of reply

#155 dachengx closed 2 years ago
2
Remove posterior charge distribution

#154 dachengx closed 2 years ago
0
Review zaq

#153 greatofdream closed 2 years ago
0
CNN might lose its self-regulared sparsity

#152 dachengx closed 2 years ago
3
Description of compound Poisson distribution

#151 dachengx closed 2 years ago
2
定义 posterior charge distribution

#150 heroxbd closed 2 years ago
0
Refining article

#149 littlewu2508 closed 2 years ago
0
Details about CNN

#148 littlewu2508 closed 2 years ago
0
Fix bug in non-Gibbs FSMP

#147 dachengx closed 2 years ago
0
Note/figures/profile.pgf 上加一个 t_0 和 mu 影响的示意

#146 heroxbd closed 2 years ago
3
把 t_ALL 和 t_1st 比较图分成两个子图

#145 heroxbd closed 2 years ago
2
Try fix mcmc & mle workflow, considering npe = 0

#144 dachengx closed 2 years ago
0
waveform shifting 和 fourier deconvolution 中的 charge 看起来过大

#143 heroxbd closed 2 years ago
3
Explain q_th; change to V_th; add V_th symbol

#142 littlewu2508 closed 2 years ago
0
按 D_w 极小改写 waveform shifting

#141 heroxbd closed 2 years ago
0
batch 新增的 MLE 部分偶然发现了运行时错误

#140 Berrysoft closed 2 years ago
5
把 \sigma_q 等改写成 Var

#139 heroxbd closed 2 years ago
0
Thanks referee

#138 dachengx closed 2 years ago
0
Specify Delta t in waveform shifting method

#137 littlewu2508 closed 2 years ago
0
Stop using boxplot

#136 dachengx closed 2 years ago
0
给 deconvolution 一个介绍。

#135 heroxbd closed 2 years ago
2
在致谢里感谢 Referee 1 的精彩评论

#134 heroxbd closed 2 years ago
0
加上 log-normal SER shape 和 gain distribution 为 gamma 的解释。

#133 heroxbd closed 2 years ago
0
Gibbs 2D MLE

#132 dachengx closed 2 years ago
0
Introduction 中 KamLAND 和几个其它实验的引用最好换成探测器硬件论文

#131 heroxbd closed 2 years ago
2
Add Double Chooz's PSD for waveform shifting method

#130 littlewu2508 closed 2 years ago
3
Symbols

#129 heroxbd closed 2 years ago
1
加入中微子实验的简介，点明 PMT 的工作模式以及为什么要研究波形分析。

#128 heroxbd closed 2 years ago
1
更多介绍中微子和暗物质实验

#127 heroxbd closed 2 years ago
1
Clarify two modes of PMT, and the role of waveform analysis

#126 littlewu2508 closed 2 years ago
0
add some text for waveform shifting method

#125 littlewu2508 closed 2 years ago
0
FSMP 在液闪发光曲线下的 t0 输出

#124 heroxbd closed 2 years ago
6
Gibbs test

#123 dachengx closed 2 years ago
1
实现 tensorflow FSMP

#122 heroxbd closed 2 years ago
1
Line 362: "the message is clear" ? I do not think so. Fig 15b is very confusing. The "ratio" is undefined. The performance of "1st" is only a few percent worse. Secondly, the result is inconsistent with the earlier Fig. 3. Only modest improvement is expected with waveform analysis and only under some circumstances according to Fig. 3.

#121 heroxbd closed 2 years ago
1
Line 324: I very much suggest the authors rewrite this paper based on only the FSMP technique and provide evidence with data that it works well. And provide much more detail on its performance for a variety of conditions.

#120 heroxbd closed 2 years ago
2
Line 238: I suggest not spending time in the paper on techniques that donot work well.

#119 heroxbd closed 2 years ago
1
Eq. 3.17: As remarked earlier, the assumption of a normal distribution for the charge gain is wrong. And it creates a problem for the model since with a normal distribution the charge can be negative which is unphysical.

#118 heroxbd closed 2 years ago
7
Fig. 9: The demonstration is just not generic enough. The pulse structure assumed is relatively compact with only a few photons and with good single PE resolution.

#117 heroxbd closed 2 years ago
2
Section 3.5.1: There some conceptual problems here. I do not see how RSS can work well with noise and with pulses that have a sparse tail of photons. I think this method is also too specific for a pulse that has relatively compact shape. For pulses that are long with sparse tails this will not work.

#116 heroxbd closed 2 years ago
1
Line 253: The referee has a hard time understanding why NPE needs to be left unknown. If NPE is large then the integral of the pulse divided by the mean gain should be a very close estimate of NPE . If NPE is small the gain fluctuation will prevent a good estimate, but it is still reasonably known.

#115 heroxbd closed 2 years ago
0
Figure 8: This uses language that is probably unfamiliar to the readers of JINST. What is a "hinge" and what are "whiskers" ? We do not do 25% and 75% confidence intervals usually.

#114 heroxbd closed 2 years ago
4