What a great opportunity for Iran!
01.03.2026 16:54
👍 1
🔁 0
💬 0
📌 0
What a great opportunity for Iran!
I cannot understand how this got through a peer reviewed process.
Where are Seq2seq, Attention mechanism… How are applied these methods to our field?
More great researchers should apply their fundamentals discoveries to applied field. (2/2)
I’m currently working on a scientific work with some interns and the gab between applied and fundamental research is insane 🤯.
In wind power forecasting there are so many papers that only mention using LSTM. Nothing more, as if there was only one was to use them. (1/2)
This life ethics is so interesting!
I would agree with most of what Pavel talked about.
I think moderations is king and that the most important thing are human connections - Aka follow with moderation for the vast majority of people listening this.
Source : youtu.be/qjPH9njnaVU?...
Before 2023: Writing perfect text was badass;
Now 2025: Writing with mistakes is badass.
If you use either BlueSky, X or LinkedIn, that would be awesome if you stop using AI to generate all your text. God I’m so tired to read those posts…
PS: anyone has suggestions for other great quantum physics foundation papers?
(6/6)
How many Einstein we lost because they listened to others and got evaluated by metrics that weren’t designed for them in the first place ?
Paper: inters.org/files/einste...
(5/6)
In 1905, he published 3 of the most influential papers in Physics.
And knowing this the quote :
“If you judge a fish by its ability to climb a tree, it will live it whole life believing that he is stupid” takes all his meaning.
(4/6)
He must have face so much critics and judgments from his family, friends, colleagues…
If you still doubt this - apply for PhDs for 4 straight years in heavy schools and after you get rejected everywhere tell everyone you are going to be all in on your research projects by yourself.
(3/6)
No professors nor universities believes in his ability to do great research - Rejected everywhere he applied from 1901 to 1905.
Still he believes in his ideas so much that he decided to go all in on them.
(2/6)
This morning I read one of the most badass paper of all time “On a Heuristic Point of View Concerning the Production and Transformation of Light” by no one else then Albert Einstein, 1905.
PhD Rejected for four straight years he lead his researches on Physics while working a full time job.
(1/6)
I have the impression that some of the leading scientists, that advocate for AI safety, also work, simultaneously, on ground breaking AI research development.
Is it just me or this is ironical ?
* Source - visit their google scholar and read their most recent papers.
With AGI it should be able to do inference on outside of distribution data. I would argue that the current performance of our best model has seen the test data somewhere in their training set - Data leakage.
I would love to hear from you on my statement! (4/4)
We optimized the model with a gradient descent methodology minimizing a selected loss function that should give us optimal weights models for that specific dataset, and that generalize well on the validation set - Similar data distribution. (3/4)
We are using generative AI algorithms trained on as many good quality data as possible. So in theory the model has as much information of the world as it has been trained on - Which is a lot! (2/4)
I think that I am missing something big. I really don’t see how we are anywhere close to AGI - Please can someone share scientific proof that we are reaching AGI any time soon ?
Let me explain my current stand on the field! (1/4)
Here is the official Artificial Intelligence for the Earth Systems journal link!
journals.ametsoc.org/view/journal...
We should create a new research direction - Maximizing researcher contributions to humanity.
I’m sure lot of great researchers have awesome ideas for that!
Why not creating a research labs with the advantages of academia & industry?
- Easy to access funds
- Project start immediately
- Full project freedom
- PI can hired whoever they want
- Good salaries
- Only paper work are manuscripts
- Open source
Is there something that makes it impossible?
I think Canada needs a clearer definition of what “top talent” means before prioritizing recruitment from abroad.
Many of us already here are struggling to find opportunities - not because of a lack of talent.
Policies like this only push young researchers to leave and do great research elsewhere.
As a beginners myself, I highly recommend this French video from one of the most exceptional professors in the domain to get started with QC: youtu.be/XkGaitu3EbI?...
I think AI Quantum Computing has tremendous potential - More people should consider doing research in that field.
Starting from first principles, creating a computer that follow Quantum Mechanics principles makes so much sense.
I see a lot of great research from that field in the coming years!
Alicia Jolicoeur-Martineau research is really fascinating!
Doing this kind of research by yourself is astonishing - I am impressed and inspired 🚀.
This bio-inspired algorithm is promising in my opinion, a must read.
arxiv.org/pdf/2510.04871?
I take a lot of pride in this paper since it was done as an undergraduate student-Nothing is impossible with hard work.
I love research, and I truly do it out of pure passion. I feel alive when collaborating with highly talented people and being around those who understand and challenge my ideas.
3.We showed that a model trained only with the predictand achieved similar performance to a model trained with multiple predictors, challenging prevailing beliefs in our field.
2.We created a model that incorporates learned interpolation directly within the network instead of performing it beforehand. By reducing the model inputs by a factor of 16 (4× downscaling), we showed that learned interpolation achieves performance equal to that of traditional interpolation.
We decided to challenge these ideas by:
1.Comparing models trained with multiple sparse and nearby domains. We showed that using sparse domains maintains performance equal-better than that of nearby domains.
We also demonstrated that adding more domains keeps performance unchanged or improves it.
My very first journal paper has officially been accepted 🤩
In this paper, my team and I presented a DL downscaling model that paves the way to Foundation Models.
Literature
1.Conducted on small geographical domains
2.Used interpolation to fit both input and output grids
3.Required lot of predictors
Nowadays, anyone with hard work, talent, and an internet connection can innovate in DL - a lot of papers and datasets are open source, and many resources are available online (GPUs, TPUs).
Collaborators will become a must at some point, but strong performance and talent are attractive and rare.
I will never retire from research.
I truly do it for fun, and I would even pay just to do it with talented people.