Closed amartinhuertas closed 8 months ago
Hi @amartinhuertas,
I've added the poster to the repo. For this poster, I use the LaTeX-templates from Pascal (the same repo as the report, he has a folder exclusively for the poster) and divide the poster into 6 sections:
The Methods section and the Results and Findings section are basically complete right now. For the introduction, only Neural network related background is presented there and I've left some space for the Geoid and Mantle Convection problem. Also, I added what was discussed during our last meeting in the Future Works section in both poster and report.
FYI, when I try to compile the poster, I get a compilation error. The error is related to a missing figure. Have you committed and pushed all the figures in the figures
folder?
Ok, I now see that the paths provided in the includegraphics
commands of your poster Latex sources are not correct (you should eliminate the poster/
prefix out of the paths). Can you please fix this? Thanks!
Hi @amartinhuertas,
Sorry for the inconvenience, I've now eliminate the poster/
prefix out of the paths.
No worries.
Some immediate comments on the poster:
Re eq (1) in the poster:
Hi @amartinhuertas,
I just pushed an updated version of the poster.
Re eq (1) in the poster:
- Use \mathcal{N} instead of N to denote the neural network as a parametric function depending on \theta (i.e., an array with all parameters)
- n-1 sub-index seems to be formated differently from n.
- What n denotes? Define it.
- What is the general form of \Theta_j (i.e., W_j x_j + b_j), sizes of W_j and b_j. Etc.
The eq (1) has been updated now and these issues should be solved, I also updated the same equation in the Report as well.
- Images in the spirit of Figs. 1, 3, 5, 9 of the 2D Mars Mantle Convection paper may significantly ease presenting the workflows you have experimented with. This applies both for the report and the poster.
I have created and added four more figures to the poster which are variations of Figs. 1, 3, 5, 9 of the 2D Mars Mantle Convection paper and the workflows look more intuitive now. I haven't added these figures to the report and I will do this later.
- Is there an expectation for a references section in the poster? If yes, you could also add a pointer to the Github repo with the codes, along with other proper references to papers, books, etc.
I have checked the Poster Specification coming out on Wattle today, and the reference section is not mentioned. Nevertheless, I still added a "Literature" section to the Poster listing the three most referenced work in the report (they are also discussed in the "Related Works" section of the report)
- The more figures and the less text the better. By now, I see quite a lot of text.
I will continue trying to remove more text from the poster to save some space for the figures.
I tried to compile the report and the poster. I get compilation errors due to missing figures. Same reason as above?
Hi @amartinhuertas,
Sorry about making the same mistake. I've removed all the redundant prefix for the path of the figures in Poster and the Report. Let me know if that still doesn't work.
Now it works. Thanks! Let us talk about it in the next meeting.
BTW, Kerl, 2022, is not actually a PhD thesis. It is a Master Thesis.
An idea which just came into my mind ... we can make the PDF of the poster to be an animated poster. You can show this animated poster on the screen of your computer, along with the printed version of the poster. More details in the next meeting.
https://www.anu.edu.au/students/academic-skills/research-writing/posters https://guides.lib.wayne.edu/posters/latex https://www.youtube.com/watch?v=2ZWnFFhVkdE https://www.youtube.com/watch?v=IDEj6irOLbQ