Mohsin retweeted
NEW: Days before Trump launched the Iran war, Trump summoned then-ODNI Tulsi Gabbard to the Oval Office.
He asked if the U.S. should go to war.
The IC assessed killing Khamenei would usher in hardliners, regime would close Strait, and Iran would attack in MidEast.
Mohsin retweeted
Poster is up. See you at 16:00 poster 15 zone B at #ISPRS2022. If you can't attend, check the video: youtu.be/StssjGAgqLk
Mohsin retweeted
I’ve been using #AnnotatedEquations in my recent papers. I think it really adds to the readability and understanding of the math.
Here are some examples. It uses #tikz in #latex.
Let me know if you like it. Happy for any feedback. [GitHub link: next tweet]
#AcademicChatter
+
Mohsin retweeted
🔎📈Let’s Transform #TimeSeries Clustering!
We designed a Noise-Aware Transformer-based #Autoencoder to cluster millions of noisy time series and retrieve the interesting temporal patterns.
Thanks to @MarcCoru @RibanaRoscher @_Syd_Ali @xiaoxiang_zhu @RenMania1 for collaboration.
Mohsin retweeted
Maybe instead of just epistemic and aleatoric uncertainty, we need a third option which measures how uncertain you are about which of these words to use
Mohsin retweeted
The paper should not only give you flashbacks to 70s music but also make you reconsider the use of temporal differences in nightlight as a proxy for local economic growth. Local growth and changes in NTLs are uncorrelated in our village dataset from Nigeria, Uganda, and Tanzania
Mohsin retweeted
There's alot of work coming out looking at finite vs infinite NNs/GPs and arguing as part of their results that finite NNs/GPs differ in that they have heavy-tailed outputs. While correct, I would argue its an impoverished view. @roydanroy @jzavatoneveth @sebnowozin @gpleiss
Mohsin retweeted
Prof. @xiaoxiang_zhu is seeking applications for a #ResearchAssociate in the market of #AI for #EarthObservation in #SmartCities. Apply #ASAP to join the @zhu_xlab at @TU_Muenchen👉portal.mytum.de/jobs/wissens… #PleaseRT @ai4eo_de @DLR_en @DLR_de @TUMPostdoc
Mohsin retweeted
Our paper "Metamorphic Testing of Datalog Engines" has been accepted and will appear at ESEC/FSE2021 @FSEconf 🥳
Preprint and code coming soon!
Mohsin retweeted
Deepfake satellite imagery poses a not-so-distant threat, warn geographers theverge.com/2021/4/27/22403…
Mohsin retweeted
Did you know you can now search scholarly materials preserved in the @InternetArchive? Meet Internet Archive Scholar, a search engine for 25M research articles from 18th C journals to #OpenAccess conf. proceedings to #Preprint publications.
blog.archive.org/2021/03/09/…
Mohsin retweeted
I have a new paper on how to represent part-whole hierarchies in neural networks.
arxiv.org/abs/2102.12627
Mohsin retweeted
How expressive are Bayesian neural nets? 🤔@AndrewFoongYK and I had different takes on this at NeurIPS this year so we're organizing a joint talk/discussion on March 11 at 17:00 GMT. Sign up for zoom link here! ✏️ forms.gle/CpeahX7a7cNS5kmz5
Mohsin retweeted
All Earth satellites orbit the globe and must observe the same spot in regular time-intervals. Why are so few people in remote sensing (measured by IGARSS reviews) working on models to capture temporal behavior? Is it the effort to use/download many images instead of just one?
Mohsin retweeted
Today is #11February & #WomenInScienceDay
We celebrate with a gallery of portraits of Women in #EUSpace & #EUDefenceIndustry
Please meet Pr. @xiaoxiang_zhu, a world-class expert in the use of #AI for #EarthObservation from @DLR_en and @TU_Muenchen 🇩🇪
Mohsin retweeted
1/6) Time for a short thread about our brain2pix paper! @lynnle_ai @artcogsys
In this paper we reconstruct natural videos (Dr. Who episodes) from the brain activity of people watching them in a fMRI scanner.
biorxiv.org/content/10.1101/…
Mohsin retweeted
Watch here to learn more about our new Munich Data Science Institute #mdsi of @TU_Muenchen youtu.be/VM7zAcBnD-A
Mohsin retweeted
Microsoft and CMU researchers begin to unravel 3 mysteries in deep learning related to ensemble, knowledge distillation & self-distillation. Discover how their work leads to the first theoretical proof with empirical evidence for ensemble in deep learning: aka.ms/AAavp1k





























