layout | comments | title | author | date |
---|---|---|---|---|
post |
true |
Post Template |
Zhenghao Peng (Team 00) |
2022-09-19 |
This block is a brief introduction of your project and must be finished. The introduction will be shown in the main page of this website.
{: class="table-of-content"}
- TOC {:toc}
Your article starts here. You can refer to the source code of lil's blogs for article structure ideas or Markdown syntax. We've provided a sample post from Lilian Weng and you can find the source code here
Please create a folder with the name of your team id under /assets/images/
, put all your images into the folder and reference the images in your main content.
You can add an image to your survey like this: ![YOLO]({{ '/assets/images/team00/object_detection.png' | relative_url }}) {: style="width: 400px; max-width: 100%;"} Fig 1. YOLO: An object detection method in computer vision [1].
Please cite the image if it is taken from other people's work.
Here is an example for creating tables, including alignment syntax.
column 1 | column 2 | |
---|---|---|
row1 | Text | Text |
row2 | Text | Text |
# This is a sample code block
import torch
print (torch.__version__)
Please use latex to generate formulas, such as:
or you can write in-text formula
You can find more Markdown syntax at this page.
Please make sure to cite properly in your work, for example:
[1] Dwibedi, Debidatta, et al. "Counting out time: Class agnostic video repetition counting in the wild." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2020.
[Peng, et al.] Peng, Zhenghao, et al. "Maybe you can also use other format for reference as you wish." Nature. 2022.
Experiment | Parameters | Results | Comments |
---|---|---|---|
DL + Data |
| Predicting only velocity | Dataset size : 10000
Network : 2->5->5->1
activation: ReLU | ~100% accurate | Generalises well over various initial velocities |
| Predicting only displacement | Dataset size : 10000
Network : 2->16->16->1
activation: ReLU | Reasonable | Better prediction for
Network : 2->16->16->2
activation: tanh | Reasonable | Better prediction for
| DL + Physics |
| Predicting both
Network : 2->16->16->1
activation: ReLU | ~0% accuracy | Expected result as no supervision of any kind is provided |
| Predicting both
Network : 2->16->16->1
activation: ReLU | Reasonable | Prediction of
Network : 2->16->16->1
activation: ReLU | Reasonable | Not a better result w.r.t direct supervision |
Observations :
- Physics equations are certain in this case and are the best to use.
- Both DL, Hybrid(DL+Physics) methods performance are equivalent (actual accuracy/loss varies based on fine training, random dataset generation)
Re running the above experiments with Dataset size of 200(Data Starvation), yielded the following observations
- DL performance is comparable with 10000 dataset when trained on much mode epochs(5x)
- Hybrid(DL+Physics) without direct supervision on
$s_t$ has comparable/better closeness than DL only method for limited epochs($\sim$300) training.
Experiment | Parameters | Results | Comments |
---|---|---|---|
DL + Data | \ | ||
Predicting both |
Dataset size : 10000 Network : 2->16->16->2 activation: tanh |
Reasonable | Better prediction for |
DL + Physics | |||
Predicting both using Loss $L_{physics} = |v_{predicted}^2-u_{initial}^2-2gs_{predicted}|$ |
Dataset size : 10000 Network : 2->16->16->1 activation: ReLU |
~0% accuracy | Expected result as no supervision of any kind is provided |
Predicting both using Loss $L_{velocity+phy} = (v_{predicted}-v_{actual})^2+\gamma*(v_{predicted}^2-u_{initial}^2-2gs_{predicted})^2$ |
Dataset size : 10000 Network : 2->16->16->1 activation: ReLU |
Reasonable | Prediction of |
Predicting both using Loss $L_{supervised+phy} = (v_{predicted}-v_{actual})^2+(s_{predicted}-s_{actual})^2+\gamma*(v_{predicted}^2-u_{initial}^2-2gs_{predicted})^2$ |
Dataset size : 10000 Network : 2->16->16->1 activation: ReLU |
Reasonable | Not a better result w.r.t direct supervision, but bettr than DL when |
Observations :
- Both DL, Hybrid(DL+Physics) methods performance are similar, Hybrid(DL+Physics) is better when
$u0$ is out of dataset, DL is better for$u0$ in dataset. - Physics equations are not certain in this case and the above methods are better to use than Physics.
- Similar observations as in data rich