# Welcome to the SMPE 2024-2025 pad
[TOC]
### Important URLs
- Pad: https://bit.ly/SMPE-2025
- [Lecture official webpage on Github](https://github.com/alegrand/SMPE/tree/master/sessions/2024_10_Grenoble)
- [The MOOC RR1 (*Reproducible research: Methodological principles for a transparent science*)](https://www.fun-mooc.fr/fr/cours/recherche-reproductible-principes-methodologiques-pour-une-science-transparente/) and [The MOOC RR2 (*Practices and tools for managing computations and data*)](https://www.fun-mooc.fr/en/courses/reproducible-research-ii-practices-and-tools-for-managing-comput/)
- [Emojis in markdown](https://gist.github.com/rxaviers/7360908)
### Communication medium
- [Mattermost](https://framateam.org/smpe-2024-2025/channels/town-square). :warning: Register through the [**Invitation link**](https://framateam.org/signup_user_complete/?id=8ixg8yt1dfna5c41mashiaxi8r&md=link&sbr=su)
### Registered Students
1. Georgios NORDQVIST
+ [Github](https://github.com/nordqvig/SMPE2024GeorgiosPublic)
+ Mattermost ID : @georgios
+ Applied Mathematics using mainly Python & Matlab
+ Erasmus student
2. Vanny Ratanak CHHEANG
+ [Github](https://github.com/vannyratanak/SMPE): https://github.com/vannyratanak/SMPE
+ Mattermost ID : @chheang_vanny_ratanak
+ Favorite tools: Python, Jupyter, Git
+ MOSIG Distributed computing: from cloud to edge computing, embedded systems and networking (DC)
3. Rintaro Ito
+ [Github](https://github.com/Rin0326/lecture)
+ Mattermost ID : @ritaro
+ Facorite tools : python, Visual Studio Code
+ Exchange Student form Japan
4. Fouad GASMI
+ [Github](https://github.com/keserz/SMPE24)
+ Mattermost ID : @gasmif
+ Facorite tools : Python, Rstudio
+ MOSIG AI4GIVR
5. Mahdi Rasouli
+ [Github](https://github.com/mehtee/SMPE_2024_ENSIMAG)
+ Mattermost ID : @mahdir
+ Facorite tools : Python, Jupyter (Google Colab/VSCode), Git
+ MOSIG DSAI
6. Demoulin Raphaël
+ [Github](https://github.com/ranDemoulin/SMPE_2024)
+ Mattermost ID : @demoulir
+ Favorite tools : C, Github
+ MOSIG AI4GIVR
7. Pedro Ernesto de Oliveira Primo
+ [Github](https://github.com/pedroernesto00/SMPE)
+ Mattermost ID: @pedroernesto
+ Favorite tools: Python, SQL, Jupyter, VS Code
+ MOSIG DSAI
8. Sarah BOUARABA
+ [Github](https://github.com/sarahbrb/SMPE_M2-MoSIG_ENSIMAG_DSAI)
+ Mattermost ID : @sarahbrb
+ Favorite tools : Python, SQL, Jupyter, Google Colab
+ MOSIG DSAI
9. Ryuhei KAWABATA
+ [Github](https://github.com/monokemonoke/SMPE_M2-MoSIG_ENSIMAG_DSAI)
+ Mattermost ID: @ryuhei-kawabata
+ Favorite tools: Python, Jupyter, Git, SQL
+ MOSIG DSAI
10. Leslie Janine Gonzalez Blancas
+ [Github](https://github.com/LeslieJanine/SMPE)
+ Mattermost ID: @leslie_gonzalez
+ Favorite tools: Python, Jupyter
+ MoSIG M2 - DSAI
+
11. José Rodolfo Mondragón-Zenteno
+ [Github](https://github.com/Moderagon97/SMPE_Mondragon)
+ Mattermost ID: @rodolfo-mondragon
+ Favorite tools: MATLAB, Python, Jupyter, VSCodium
+ Research Area : Machine Learning and Robotics
+ MSc Student: MoSIG M2 - AI4GIVR
+ Homeworks progress: :D
+ R & RStudio Installed (Jupyter notebook in Kaggle)
+ MOOC registration: ✔
+ Start learning R: in progress
+ MOOC modules:
- Module 1: in progress
- Module 2: pending
12. Serban SCORTEANU
+ [Github] (https://github.com/sserban94/SMPE_2024_Ensimag)
+ Mattermost ID: @serban.scorteanu
+ Favorite tools: JVM languages, Python, Git, PostgreSQL, Intellij/Vscode
+ MOSIG M2 Double Degree - SHCE
13. PHAN Manh Tung
+ [Github] (https://github.com/phanmanhtung/SMPE-MoSIG-M2)
+ Mattermost ID: @phantu
+ Favorite tools: Python, Git, Google Colab
+ MOSIG M2
# Lecture 1 : Introduction (Arnaud)
## Arnaud Legrand: Lecture Organization
- Python : 13
- R: 1
- Notebooks (Rstudio , Jupyter , Org-Mode ): 12
- Git (, Annex ): 11 + 0
- Zenodo, SWH: 0
- Docker : 3 + 1
- Confidence Interval : 6
- Credibility Region : 0
- Observation vs. Experiment: 1 HARKING
- P-value, P-hacking (): 1
- Linear regression (): 8
- Model Selection (): 2
- Design of Experiments (): 1
- Reinforcement Learning ():
## Arnaud Legrand: The Reproducible Research *Crisis*/Movement
## Introduction to R and the tidyverse (dplyr and ggplot2)
# Lecture 2: Data visuasation and epistemology of computer science (Jean-Marc)
It seems like none took any note... :shrug:
# Lecture 3: Conducting Experiments (Arnaud and Céline)
## Feedback
1. Georgios NORDQVIST ([Github](https://github.com/nordqvig/SMPE2024GeorgiosPublic))
+ Good organisation of your repos. :clap:
+ Good summary of Popper's text.
+ Well done for the critical analysis of Jean-Marc's graphs although I do not agree with all of them.
3. Vanny Ratanak CHHEANG ([Github](https://github.com/vannyratanak/SMPE)
+ :construction:
4. Rintaro Ito ([Github](https://github.com/Rin0326/lecture))
+ :construction:
5. Fouad GASMI ([Github](https://github.com/keserz/SMPE24))
+ Empty
6. Mahdi Rasouli ([Github](https://github.com/mehtee/SMPE_2024_ENSIMAG)
+ :construction:
7. Demoulin Raphaël ([Github](https://github.com/ranDemoulin/SMPE_2024))
+ Empty
8. Pedro Ernesto de Oliveira Primo ([Github](https://github.com/pedroernesto00/SMPE))
+ Good reading of Popper's text. I recommand using plain text or markdown instead of pdf for such document though.
9. Sarah BOUARABA ([Github](https://github.com/sarahbrb/SMPE_M2-MoSIG_ENSIMAG_DSAI)
+ :construction:
10. Ryuhei KAWABATA ([Github](https://github.com/monokemonoke/SMPE_M2-MoSIG_ENSIMAG_DSAI))
+ Empty
11. Leslie Janine Gonzalez Blancas ([Github](https://github.com/LeslieJanine/SMPE))
+ :construction:
## Quicksort
### Issues
- :shocked_face_with_exploding_head: if you're on windows
- Requires a C compiler, perl, R, gnuplot
- Filenames contain `:`
- Compiling on Mac and Windows required removing the `-lrt` option
- But we now have experiments for Windows (with WSL), 2 Mac, and 2 Ubuntu :ok:
- The `R` plot was broken (colors did not show) because the behavior of the `read.csv` command changed (the `Type` column whose values are `Sequential`, `Parallel`, `Built-in` is now a vector of strings instead of a vector of factors). This was easily fixed by adding a `as.factor(df$Type)`.
- 
- 
- `ggplot(df, aes(x=Size, y=Time, color=Type)) + geom_point() + geom_line()`
- 
- ```
df %>% group_by(Size, Type) %>% summarise(MeanTime=mean(Time)) %>%
ggplot(aes(x=Size, y=MeanTime, color=Type)) + geom_point() + geom_line() + theme_bw()
```
- 
## Empirical evaluation of Fitts’s law
### Laboratory Notebook for a user pointing experiment
- 
- 
- 
- 
- Code to plot both:
- ```
{r Plotting the *mean data together with the linear regression}
ggplot(meanMTdf, aes(x=ID, y=MT,color=origen, shape=origen)) +
geom_point() +
geom_smooth(method='lm')
```
- An other experiment using the same parameters of the journal:
- Cornell page results:
- 
- Results form the notebook
- 
- Comparing the original data to the new one
- 
- Code to plot 2 graphs without binding dataframes (this is not recommended though, binding data-frames while adding an `Origin` column is much cleaner):
```{r}
p <- ggplot() +
#plot 1
geom_point(data=meanMTdf_o,aes(x=ID, y=MT)) +
geom_smooth(data=meanMTdf_o,method='lm',aes(x=ID, y=MT), fill="blue",colour="darkblue", linewidth=1) +
#plot 2
geom_point(data=meanMTdf,aes(x=ID, y=MT)) +
geom_smooth(data=meanMTdf,method='lm',aes(x=ID, y=MT), fill="red",colour="red", linewidth=1)
print(p)
```
# Lecture 4: Data management (Arnaud and Céline)
## Feedback
1. Georgios NORDQVIST ([Github](https://github.com/nordqvig/SMPE2024GeorgiosPublic))
+ [x] Repos organisation. :clap:
+ Beware, your naming conventions will soon make things difficult to manage. More about this today. :wink:
+ [x] Summary of Popper's text. :clap:
+ [x] Critical analysis of Jean-Marc's
+ You're too kind. Imho, the figures really fail most of the checks.
+ :warning: links to images #2 and #3 in your md file are broken.
+ [ ] Challenger :rocket:
+ [x] Pointing/Quicksort
+ New data, new analysis, detailed comments on what has been done. :clap:
+ :warning: Pushing the output could be helpful for the reader.
+ [ ] Pointing/Quicksort improvement
2. Vanny Ratanak CHHEANG ([Github](https://github.com/vannyratanak/SMPE)
+ Nothing new :construction:
3. Rintaro Ito ([Github](https://github.com/Rin0326/lecture))
+ Nothing new :construction:
4. Fouad GASMI ([Github](https://github.com/keserz/SMPE24))
+ [x] Summary of Popper's text. :clap:
5. Mahdi Rasouli ([Github](https://github.com/mehtee/SMPE_2024_ENSIMAG)
+ :construction:
6. Demoulin Raphaël ([Github](https://github.com/ranDemoulin/SMPE_2024))
+ Empty
7. Pedro Ernesto de Oliveira Primo ([Github](https://github.com/pedroernesto00/SMPE))
+ [x] Repos organisation
+ [x] Summary of Popper's text
+ Good reading of Popper's text. I recommand using plain text or markdown instead of pdf for such document though.
+ [ ] Critical analysis of Jean-Marc's
+ [ ] Challenger :rocket:
+ [ ] Pointing/Quicksort
+ [ ] Pointing/Quicksort improvement
8. Sarah BOUARABA ([Github](https://github.com/sarahbrb/SMPE_M2-MoSIG_ENSIMAG_DSAI)
+ [x] Repos organisation :clap:
+ [x] Summary of Popper's text :clap:
+ [x] Critical analysis of Jean-Marc's :clap:
+ Unimportant point: link to Jean-Marc's slides on "What is science" is broken. I have to say I wasn't expecting anyone to criticize the figure from this set of slides. I'll forward it to Jean-Marc. :laughing:
+ "Une brève histoire de l'informatique" Figure: I agree with all your critics and suggestions. Note that the fact that lot of data crammed is probably a design choice (i.e. the message is "See how much things have happened"), but it's a poor excuse for being unreadable.
+ "Banana": agree with the critics. Try to follow the checklist to make sure you check everything. Are you sure the 2D bar chart would solve the representation issues of such kind of data ?
+ "Climate": Agree. In general, just like perspective, we do not all have the same perception of colors (without even talking about b&w printing) so lineplots with grids are generally better, even though color may also be used to play on "emotions" in this case.
+ "Finance and crypto-currency": I obviously know less about finance than you do. :stuck_out_tongue: Good criticisms.
+ [x] Challenger :rocket: :clap:
+ Excellent. You caught everything that was to see. My only suggestions would be to try to add the uncertainty (confidence region) to the logistic regression plot.
+ Not sure Ridge regression or Lasso regression would make much sense here with so few parameters. I fear you do not have enough data to investigate non-linear models or Mixed effects model.
+ [x] Pointing/Quicksort
+ [ ] Pointing/Quicksort improvement
+ Can't wait to read your new notebooks. :wink:
9. Ryuhei KAWABATA ([Github](https://github.com/monokemonoke/SMPE_M2-MoSIG_ENSIMAG_DSAI))
+ Empty
10. Leslie Janine Gonzalez Blancas ([Github](https://github.com/LeslieJanine/SMPE))
+ :construction:
11. José Rodolfo Mondragón-Zenteno ([Github](https://github.com/Moderagon97/SMPE_Mondragon))
+ :construction:
12. Serban SCORTEANU ([Github](https://github.com/sserban94/SMPE_2024_Ensimag))
+ Empty
13. PHAN Manh Tung ([Github](https://github.com/phanmanhtung/SMPE-MoSIG-M2))
+ Empty
# Lecture 5: Probabilities, CI | Correlation, Causality (Arnaud)
## Feedback
1. Georgios NORDQVIST ([Github](https://github.com/nordqvig/SMPE2024GeorgiosPublic))
+ [x] Repos organisation and naming convention :clap:
+ Beware of accents, they're a pain for scripts and you still have a few ones left.
+ Having Rmd is great but I think saving the corresponding PDF or HTML files is a good thing as rerunning with the same environemnt may be hard for others.
+ MOOC.
+ I like your notes on the MOOC
+ [ ] Challenger :rocket:
2. Vanny Ratanak CHHEANG ([Github](https://github.com/vannyratanak/SMPE)
+ Nothing new :construction:
3. Rintaro Ito ([Github](https://github.com/Rin0326/lecture))
+ Nothing new :construction:
4. Fouad GASMI ([Github](https://github.com/keserz/SMPE24))
+ [x] Summary of Popper's text. :clap:
+ [-] Pointing Experiment (early impressions, but no new data nor improvements of the original project)
+ [-] Analysis of graphs (Ongoing)
+ I disagree with the origin choice. Starting the Y-axis at 0 would probably greatly change the impression of the reader. I see no reason for not starting at 0.
5. Mahdi Rasouli ([Github](https://github.com/mehtee/SMPE_2024_ENSIMAG)
+ Nothing new :construction:
6. Demoulin Raphaël ([Github](https://github.com/ranDemoulin/SMPE_2024))
+ Still Empty
7. Pedro Ernesto de Oliveira Primo ([Github](https://github.com/pedroernesto00/SMPE))
+ [x] Repos organisation :clap:
+ :warning: you still have spaces (and `'`) in your filenames, which is painful when scripting. Use `_` instead.
+ [x] Summary of Popper's text :clap:
+ [x] Critical analysis of Jean-Marc's graphics
+ You got the idea, although I would be harsher than you in my critics. :smiling_imp:
+ [ ] Challenger :rocket:
+ [x] Quicksort improvement
+ New data has been generated and a new notebook has been created.
+ Write more comments though on what you get from all this!
8. Sarah BOUARABA ([Github](https://github.com/sarahbrb/SMPE_M2-MoSIG_ENSIMAG_DSAI)
+ [x] Repos organisation :clap:
+ Your renamed all your files, that's good. You still have a few spaces in some of your filenames/directory but that's OK.
+ [x] Summary of Popper's text :clap:
+ [x] Critical analysis of Jean-Marc's figures :clap:
+ [x] Challenger :rocket: :clap:
+ [x] Pointing
+ [x] Pointing improvement
+ I see you've starting doing linear regression analysis. That's good. No feedback from me for the moment, we'll rediscuss after the Linear Regression lecture.
9. Ryuhei KAWABATA ([Github](https://github.com/monokemonoke/SMPE_M2-MoSIG_ENSIMAG_DSAI))
+ :construction: mostly notes on the lecture at the moment
+ [x] Quicksort Experiment
+ I've seen your journal, that's good.
10. Leslie Janine Gonzalez Blancas ([Github](https://github.com/LeslieJanine/SMPE))
+ [-] Critical analysis of Jean-Marc's figures. Wow, have you redone the banana figure by yourself ?
11. José Rodolfo Mondragón-Zenteno ([Github](https://github.com/Moderagon97/SMPE_Mondragon))
+ [x] Chritical analysis of Jean-Marc's figures :clap:
+ Good work. You got it.
+ I think you should use md instead of pdf for that kind of work though.
+ [x] Popper :clap:
+ Same as above. Try to be more synthetic if you can.
12. Serban SCORTEANU ([Github](https://github.com/sserban94/SMPE_2024_Ensimag))
+ Empty
13. PHAN Manh Tung ([Github](https://github.com/phanmanhtung/SMPE-MoSIG-M2))
+ Good notes on the lecture.
## Lecture notes
# Lecture 6: Linear Regression (JMV) | Scientific integrity (Cyril)
## Feedback
1. Georgios NORDQVIST ([Github](https://github.com/nordqvig/SMPE2024GeorgiosPublic))
+ MOOC.
+ [ ] Challenger :rocket:
+ [x] Confidence intervals
+ Your computation seems complicated, especially the one with `replication_times = c(5, 5, 5, 5, 5)`. You probably missed the `n()` function which can be used in `summarise` to obtain the number of rows in the group.
+ Furthermore, there is no need to compute the sum_of_squares. Simply call the `var` and `sd` functions.
+ More important, tt seems to me you forgot to divide your sd by `sqrt(n)`.
+ All this can be done in one step with `summarise(mean=mean(Time), n=n(), err=sd(Time)/sqrt(n), lb=mean-2*err, ub=mean-2*err)` instead of extending the dataframes manually.
2. Vanny Ratanak CHHEANG ([Github](https://github.com/vannyratanak/SMPE)
+ Nothing new :construction:
3. Rintaro Ito ([Github](https://github.com/Rin0326/lecture))
+ Nothing new :construction:
4. Fouad GASMI ([Github](https://github.com/keserz/SMPE24))
+ Nothing new
+ [x] Summary of Popper's text. :clap:
+ [-] Pointing Experiment (early impressions, but no new data nor improvements of the original project)
+ [-] Analysis of graphs (Ongoing)
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
5. Mahdi Rasouli ([Github](https://github.com/mehtee/SMPE_2024_ENSIMAG)
+ Nothing new :construction:
6. Demoulin Raphaël ([Github](https://github.com/ranDemoulin/SMPE_2024))
+ Still Empty
7. Pedro Ernesto de Oliveira Primo ([Github](https://github.com/pedroernesto00/SMPE))
+ Nothing new.
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
8. Sarah BOUARABA ([Github](https://github.com/sarahbrb/SMPE_M2-MoSIG_ENSIMAG_DSAI))
+ [x] Pointing improvement
+ I see you've starting doing linear regression analysis. That's good. No feedback from me for the moment, we'll rediscuss after the Linear Regression lecture.
+ [x] Challenger :rocket:
+ [x] Confidence interval
+ Excellent work overall, as usual. This being said, here are a few comments/issues.
+ When computing CI, grouping the measurements for all different sizes does not make much sense as we know it has a strong impact and leads to very different orders of magnitude. This does not correspond with the hypothesis of the CLT.
+ You've used the Student distribution, which compensates for the the low number of samples, but assumes normality of the samples, which does not hold at all. So such computation only appears more reliable.
+ I know you probably know about all this but since you did not write it as a precaution to consider in your comments, I thought it'd be better to make sure we're on the same line.
+ [x] linear model of quicksort data
+ Indeed, there is no reason from looking at the data to consider anything more ellaborate than a simple linear model. This being said, it is clear for the parallel version that there is a problem/bias for 0.1e6. This could be a measurement problem or a modeling problem, especially as the expected is rather n.log(n)+n+log(n). Obviously, the experiment design does not allow to investigate this but if you zoom on small values you'll see for yourself that the model does not work that well.
9. Ryuhei KAWABATA ([Github](https://github.com/monokemonoke/SMPE_M2-MoSIG_ENSIMAG_DSAI))
+ Nothing new
+ :construction: mostly notes on the lecture at the moment
+ [x] Quicksort Experiment
+ I've seen your journal, that's good.
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
10. Leslie Janine Gonzalez Blancas ([Github](https://github.com/LeslieJanine/SMPE))
+ Nothing new
+ [-] Critical analysis of Jean-Marc's figures. Wow, have you redone the banana figure by yourself ? [-] Oh no, i cannot take credit for such a good improved figure. Actually, I add the reference of the site where i found it when i was looking for inspiration. [Ggplot2 effective visualization](https://pacha.dev/blog/2022/12/21/banana-exports-1994-2005/index.html)
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
11. José Rodolfo Mondragón-Zenteno ([Github](https://github.com/Moderagon97/SMPE_Mondragon))
+ Nothing new
+ [x] Chritical analysis of Jean-Marc's figures :clap:
+ Good work. You got it.
+ I think you should use md instead of pdf for that kind of work though.
+ [x] Popper :clap:
+ Same as above. Try to be more synthetic if you can.
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
12. Serban SCORTEANU ([Github](https://github.com/sserban94/SMPE_2024_Ensimag))
+ Empty
13. PHAN Manh Tung ([Github](https://github.com/phanmanhtung/SMPE-MoSIG-M2))
+ Good notes on the lecture.
+ [ ] Challenger :rocket:
+ [x] Confidence interval
+ Your approach with python and Student's law is exactly the same as the one of Sarah BOUARABA so please see my comments above.
+ I also encourage you to use a notebook where you may mix comments and code together instead of having separated (possibly unrelated) files.
14. Let's read: https://cloud.univ-grenoble-alpes.fr/s/6E5ryKcTF8jj5RS
# Lecture 7: Linear Regression + DoE (AL)
# Lecture 8: DoE (AL) + Scientific Integrity (CL)
## Feedback
1. Georgios NORDQVIST ([Github](https://github.com/nordqvig/SMPE2024GeorgiosPublic))
+ MOOC.
+ [ ] Challenger :rocket:
+ [ ] Peer evaluation
+ [ ] CI and linear regression for quicksort
2. Vanny Ratanak CHHEANG ([Github](https://github.com/vannyratanak/SMPE)
+ Nothing new :construction:
3. Rintaro Ito ([Github](https://github.com/Rin0326/lecture))
+ Nothing new :construction:
4. Fouad GASMI ([Github](https://github.com/keserz/SMPE24))
+ Nothing new
5. Mahdi Rasouli ([Github](https://github.com/mehtee/SMPE_2024_ENSIMAG)
+ Nothing new :construction:
6. Demoulin Raphaël ([Github](https://github.com/ranDemoulin/SMPE_2024))
+ Some commits with documents on visualization and Popper
7. Pedro Ernesto de Oliveira Primo ([Github](https://github.com/pedroernesto00/SMPE))
+ [x] Filenaming concentions
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
8. Sarah BOUARABA ([Github](https://github.com/sarahbrb/SMPE_M2-MoSIG_ENSIMAG_DSAI))
+ [x] linear model of quicksort data
+ See previous comment.
9. Ryuhei KAWABATA ([Github](https://github.com/monokemonoke/SMPE_M2-MoSIG_ENSIMAG_DSAI))
+ Nothing new
+ :construction: mostly notes on the lecture at the moment
+ [x] Quicksort Experiment
+ I've seen your journal, that's good.
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
10. Leslie Janine Gonzalez Blancas ([Github](https://github.com/LeslieJanine/SMPE))
+ Nothing new
+ [-] Critical analysis of Jean-Marc's figures. Wow, have you redone the banana figure by yourself ? [-] Oh no, i cannot take credit for such a good improved figure. Actually, I add the reference of the site where i found it when i was looking for inspiration. [Ggplot2 effective visualization](https://pacha.dev/blog/2022/12/21/banana-exports-1994-2005/index.html)
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
11. José Rodolfo Mondragón-Zenteno ([Github](https://github.com/Moderagon97/SMPE_Mondragon))
+ Nothing new
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
12. Serban SCORTEANU ([Github](https://github.com/sserban94/SMPE_2024_Ensimag))
+ Empty
13. PHAN Manh Tung ([Github](https://github.com/phanmanhtung/SMPE-MoSIG-M2))
+ Nothing new
+ [ ] Challenger :rocket:
## Homeworks review
Let's read: https://cloud.univ-grenoble-alpes.fr/s/6E5ryKcTF8jj5RS
- Which articles have you read ?
- The anatomy of a large-scale hypertextual Web search engine
- There was a world with search engines before Google! I also realized there were public and private search engines.
- This is the article that present the PageRank algorithm.
- The authors Sergy Brin and Lauwrance Page are actually the founders of Google. This is a reprint in a proper journal of the original article. This is why the authors present also new content on the architecture of the system.
- Attention is All you Need
- It's ML so I had read it before. It's a breakthrough in computer vision.
- It explains how the model learns places that are worth of interest, which could not be done with previous models. They explain how the feeding of blah and blah allows to make the difference.
- Is it published in a Journal ? It's from a conference actually, although there is a version on Arxiv. It was released at the same time on arxiv as in the conference, and the conference status gave it more credit but it's not so true anymore.
- All the authors are from Google and they first cite all the previous types of NN (RecurrentNN, GatedNN, etc.) to distinguish what's new from what existed before.
- On the Refinements of DHTs
- It was published in a journal.
- Tried to read it but did not understand much and could not make sense of it.
- It's actually a fake paper with no meaning. It's good English but it is meaningless.
- Take away message: it's not because it's published that it's good.
- The publisher is a small publisher just trying to make money. But there has also been papers published in IEEE and Nature.
- Wait, aren't you supposed to present the paper in a conference?
- There are also fake conferences with sessions full of such papers. It's likely the conference did not even take place.
- Deep Latent Mixture Model for Recommendation
- This one again is generated, even though there seems to be a lot of nice looking but non-sense maths.
- There is not even a conclusion
- The references are also often made up
- It is on on Arxiv but it has not been published elsewhere.
- It's probably a mix between stuff written by real people plus some generated maths.
- A Relational Model of Data for Large Shared Data Banks
- Historical paper on BNF forms.
- The founder of Oracle claims that this is the paper that inspired him.
- Chord: A Scalbale P22 Lookup Protocol for Internet Applications
- It's about DHT and P2P systems. Quite a famous (good!) one, even though there were a few glitches/mistakes in there and it generated a lot of subsequent work.
- Enhencing Sentiment analysis through Text Analysis and Data Mining.
- Again, it's published. "Again", in IEEE.
- There is something weird: Baking Vector Maching (SVM), Normal Language Handling (NLP), Comvolution of repetitive brain organizations (CNN) etc. Actually a lot of weird phrasing because it was made with a paraphrasing tool which generates tortured phrases (polylexical expressions).
Most of the stuff you'll find makes sense and is sound but sometimes, there is weird and unethical stuff happening.
## Lecture
- SciGen was designed to test the (lack of) seriousness of a conference.
- Google Scholar is full of garbage, especially for people with common names.
- Journals impersonating real existing journals in an open acces model.
- Mechanisms to detect scigen papers
- Plagiary checkers fail (after all, that's not what they're designed for) as they have very small fractions of text in common.
- Discussion on what should be found in a related work section, and why it's very different from plagiarism.
# Lecture 9: DoE (AL) + Bayesian statistics
## Feedback
1. Georgios NORDQVIST ([Github](https://github.com/nordqvig/SMPE2024GeorgiosPublic))
+ MOOC.
+ [x] Challenger :rocket: :+1:
+ [ ] Peer evaluation
+ [ ] CI and linear regression for quicksort
2. Vanny Ratanak CHHEANG ([Github](https://github.com/vannyratanak/SMPE)
+ Nothing new :construction:
3. Rintaro Ito ([Github](https://github.com/Rin0326/lecture))
+ Nothing new :construction:
4. Fouad GASMI ([Github](https://github.com/keserz/SMPE24))
+ Nothing new
5. Mahdi Rasouli ([Github](https://github.com/mehtee/SMPE_2024_ENSIMAG)
+ Nothing new :construction:
6. Demoulin Raphaël ([Github](https://github.com/ranDemoulin/SMPE_2024))
+ Some commits with documents on visualization and Popper
+ Nothing new :construction:
7. Pedro Ernesto de Oliveira Primo ([Github](https://github.com/pedroernesto00/SMPE))
+ [ ] Challenger :rocket:
+ [x] Confidence interval
+ See comments on Sarah BOUARABA's work on line 366
8. Sarah BOUARABA ([Github](https://github.com/sarahbrb/SMPE_M2-MoSIG_ENSIMAG_DSAI))
+ Nothing new this week. No worry, it's fine, you were ahead anyway.
9. Ryuhei KAWABATA ([Github](https://github.com/monokemonoke/SMPE_M2-MoSIG_ENSIMAG_DSAI))
+ Nothing new this week
+ :construction: mostly notes on the lecture at the moment
+ [x] Quicksort Experiment
+ I've seen your journal, that's good.
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
10. Leslie Janine Gonzalez Blancas ([Github](https://github.com/LeslieJanine/SMPE))
+ Nothing new this week
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
11. José Rodolfo Mondragón-Zenteno ([Github](https://github.com/Moderagon97/SMPE_Mondragon))
+ Nothing new this week
+ [ ] Challenger :rocket:
+ [ ] Confidence interval
12. Serban SCORTEANU ([Github](https://github.com/sserban94/SMPE_2024_Ensimag))
+ Empty
13. PHAN Manh Tung ([Github](https://github.com/phanmanhtung/SMPE-MoSIG-M2))
+ [ ] Challenger :rocket:
+ [x] Confidence interval
+ See comments on Sarah BOUARABA's work on line 366
+ [x] Linear regression for quicksort
+ Be careful, the train/test split approach is not at all what I showed during the lecture. What's important for our lecture is the evaluation of the confidence and the residuals analysis.
# Lecture 11: Discussions about Ethical aspects of research
Presence:
1.- José Rodolfo Mondragón Zenteno
2.- Raphaël Demoulin
3.- Gasmi Fouad
4.- Leslie Janine Gonzalez Blancas
5.- Sarah BOUARABA
6.- PHAN Manh Tung
# Lecture 12: Presentation
Here is a random student/topic assignment (you may trade a topic with someone else if you’re uncomfortable with the one you’ve been assigned; you may even pick an other topic if you can’t trade yours but I’d like everyone to work on different topics as much as posible). I have only quickly read through most of these documents so do not expect extraordinay content. Prepare a short presentation with the pros and cons of each technology. You will present (1) the problems it tries to address, (2) to what extent it does, (3) whether it unlocked other usages/problems. (4) Indicate in which of the 4th scenarios of the ADEME (https://transitions2050.ademe.fr/en) this technology is rooted.
1. Georgios NORDQVIST ([Github](https://github.com/nordqvig/SMPE2024GeorgiosPublic))
- Photo Enhancement: https://cacm.acm.org/magazines/2021/11/256376-filtering-for-beauty/fulltext
2. Vanny Ratanak CHHEANG ([Github](https://github.com/vannyratanak/SMPE))
- Affordance++: allowing objects to communicate dynamic use: http://plopes.org/wp-content/uploads/papers/2015-CHI-AffordanceLopes.pdf
3. Rintaro Ito ([Github](https://github.com/Rin0326/lecture))
- Pattern Recognition
+ https://www.researchgate.net/publication/221004855_Pattern_Recognition_in_Medical_Diagnosis_Prognosis_and_Treatment (just the abstract)
+ https://link.springer.com/article/10.1007/s13167-017-0083-9 (mostly abstract/introduction/conclusion)
+ https://www.forbes.com/sites/forbestechcouncil/2022/03/15/pattern-recognition-power-three-reasons-ai-will-improve-clinical-care/?sh=577dbe8365e3
4. Fouad GASMI ([Github](https://github.com/keserz/SMPE24))
- Mental illness detection: https://pmc.ncbi.nlm.nih.gov/articles/PMC9914523
5. Mahdi Rasouli ([Github](https://github.com/mehtee/SMPE_2024_ENSIMAG)
- Scientific article reading and writing
+ https://arxiv.org/abs/2211.09085 (Galactica)
+ https://voicebot.ai/2022/11/22/meta-halts-academic-paper-generator-ai-demo-after-3-days/
+ https://arxiv.org/abs/2408.06292 (The AI scientist, a very recent article on the topic)
6. Demoulin Raphaël ([Github](https://github.com/ranDemoulin/SMPE_2024))
- Game theory and AI/ML
+ https://medium.com/@enriqueavila.finance/the-convergence-of-ai-and-game-theory-revolutionizing-strategic-decision-making-91d47695aeda
+ https://link.springer.com/article/10.1007/s11042-022-12153-2 (if you really want to know more about how this works but the article is long and deep)
7. **Pedro Ernesto de Oliveira Primo** ([Github](https://github.com/pedroernesto00/SMPE))
- Formal methods and AI for laws
+ https://aair-lab.github.io/aia2024/papers/piskac_aia24.pdf
+ https://ceur-ws.org/Vol-1844/10000524.pdf
8. Sarah BOUARABA ([Github](https://github.com/sarahbrb/SMPE_M2-MoSIG_ENSIMAG_DSAI))
- Smart Buildings
+ https://www.propmodo.com/can-a-building-have-empathy/
+ https://www.propmodo.com/recent-facebook-hack-highlights-the-vulnerability-of-smart-buildings/
9. Ryuhei KAWABATA ([Github](https://github.com/monokemonoke/SMPE_M2-MoSIG_ENSIMAG_DSAI))
- AI and bias in decision making:
+ https://www.bbc.com/news/business-50432634
+ https://qz.com/1427621/companies-are-on-the-hook-if-their-hiring-algorithms-are-biased/
10. Leslie Janine Gonzalez Blancas ([Github](https://github.com/LeslieJanine/SMPE))
- Robotic Nurses:
+ https://www.frontiersin.org/articles/10.3389/frobt.2022.832248/full
+ https://www.online-sciences.com/robotics/healthcare-robotics-nursing-care-robots-review-types-advantages-disadvantages-uses/
+ [Slides](https://docs.google.com/presentation/d/1A8OO_sdlKLEiqYL723ApCIJm-oXWhC4x0fvDDsSgA5M/edit?usp=sharing)
11. **José Rodolfo Mondragón-Zenteno** ([Github](https://github.com/Moderagon97/SMPE_Mondragon))
- Computer Brain Interface
+ https://en.wikipedia.org/wiki/Neuralink
+ https://www.inria.fr/en/brain-computer-interfaces-nerv-multidisciplinary-project-team-seeking-help-people-paralysis
12. Serban SCORTEANU ([Github](https://github.com/sserban94/SMPE_2024_Ensimag))
- Digital Agriculture for Small-Scale Producers: Challenges and Opportunities: https://cacm.acm.org/magazines/2021/12/256930-digital-agriculture-for-small-scale-producers/fulltext
13. PHAN Manh Tung ([Github](https://github.com/phanmanhtung/SMPE-MoSIG-M2))
- Autonomous cars
+ https://www.thalesgroup.com/en/markets/digital-identity-and-security/iot/magazine/7-benefits-autonomous-cars
+ https://www.forbes.com/sites/forbestechcouncil/2022/02/14/autonomous-vehicles-and-their-impact-on-the-economy/
+ https://medium.com/@teraki/energy-consumption-required-by-edge-computing-reduces-a-autonomous-cars-mileage-with-up-to-30-46b6764ea1b7