Niharika S. D'Souza

Research Scientist,@IBMResearch | PhD @JHU_ECE + MSE @JohnsHopkinsAMS | BTech. EE @IITKGP

GNNs, Connectomics, Multimodal/Structured Representations, AI4Science

[Chhavi Yadav @chhaviyadav_]

We should totally not submit half-baked papers just to ‘get reviewer feedback’, really need to get rid of this attitude. It wastes everyone’s time & burdens an already overloaded system. For instance, I’ve reviewed 10 papers since the last month & only 2/10 are worth being (& will be) accepted, the others are just half-baked papers, which literally anyone with some sensibility would reject!
If the work isn’t ready for a conference, consider a workshop instead, and that should also be submitted with reasonable standards ONLY :(

[GRAIL workshop @ MICCAI 2026 @GRAIL2025]

🎉 Exciting News! 🎉

GRAIL 2025, the 7th International Workshop on Graphs in Biomedical Image Analysis, has been officially accepted as a MICCAI 2025 workshop! 🚀

We'll be back in-person in Daejeon, Republic of Korea for an incredible event🌏💡

[Anthony Bonato @Anthony_Bonato]

Not everyone will get this

[Physics In History @PhysInHistory]

“God used beautiful mathematics in creating the world.”

— Paul Dirac (1902 - 1984)

[UniReps @unireps]

Coming right up - "Modern Hopfield Networks meet Encoded Neural Representations - Addressing Practical Considerations" by Tanveer Syeda-Mahmood from IBM Research ✨

[Peyman Milanfar @docmilanfar]

How are Kernel Smoothing in statistics, Data-Adaptive Filters in image processing, and Attention in Machine Learning related?

My goal is not to argue who should get credit for what, but to show a progression of closely related ideas over time and across neighboring fields.

1/n

[Peyman Milanfar @docmilanfar]

The simplest things in statistics can be deceiving. Take a measurement

𝐲 = θ + σ e

where e ~ 𝒩(0,𝐈) is d-dim. What’s a good estimate of θ? You’d think the maximum-likelihood would be best, but you’d be wrong.

James-Stein dominates ML in the mean² sense for d ≥3

Stein's estimator "shrinks" the ML estimate towards the origin - which seems very counterintuitive -why move away from the data?

But by shrinking, we reduce variance significantly -and this more than compensates for the small bias introduced by shrinkage, leading to lower MSE

[Bartek Papiez @bwpapiez]

Travelling for @MICCAI_Society conference?
➡️Programme for #GRAIL2024 is online, see here: https://t.co/ExXMlxc8yL
➡️6th October morning, Morocco
➡️ fantastic talks on current research on #geometricdeeplearning for medical image analysis!
➡️fantastic keynote speakers!

[Favour Nerrise 🇨🇲 @fnerrise]

🤩 Come join us talk about two of my favourite things—the brain 🧠 and graphs 📈.

No expertise in any of those topics? No worries, I’ll break things down. Will also share some insights using neuroAI models for my research studies.

See you Aug. 30 at 9am PT! Link below.

[GRAIL workshop @ MICCAI 2026 @GRAIL2025]

🚀Exciting announcement🚀

We are launching the GRAIL Workshop Journal Club!

Our aims are to:
📚Read and discuss the latest cutting-edge papers
☕️Bring the Graph Deep Learning community closer together
🧑‍💼Create a platform to share insights, discuss, and receive research feedback

[GRAIL workshop @ MICCAI 2026 @GRAIL2025]

As the final submission deadline approaches, we are thrilled to announce two of our guest speakers for #GRAIL2024:

[Satyananda Kashyap @follow_kashyap]

📢 Deadline Extended: ML-CDS Workshop @MICCAI_Society
Submit your innovative #AI/#ML papers for clinical decision support by June 29, 2024 (AoE).
Join us in Marrakesh to shape the future of #healthcare!
Details: https://t.co/c2LHbyzKN1…

#MICCAI2024 #ClinicalDecisionSupport

[Niharika S. D'Souza @SdNiharika]

Apologies for the miscommunication earlier. Please note the latest submission deadline for GRAIL - June 29th, 23:59 PST.

We look forward to learning about your work!

[GRAIL workshop @ MICCAI 2026 @GRAIL2025]

📢Urgent: Change in Submission Deadline📢

👉The submission date is the 29th June, 23.59 PST
👉The submission portal: https://t.co/b0v5ZjLt11

Apologies for any inconveniences and confusion, we look forward to your submissions!

[Neelesh Salian 💻 @nssalian]

People using Jupyter notebooks

[Bartek Papiez @bwpapiez]

6th Workshop on GRaphs in biomedicAl Image anaLysis!

🔔20 June 2024: Deadline for Intention to submit
🕒24 June 2024: Submission deadline, 23:59 pm, PST

@GRAIL2024: @medai_ahmadi @AneesBKazi1 @kr_zaripova @SdNiharika @godson_lucy @youyegn
@MICCAI_Society @MiccaiStudents

[Niharika S. D'Souza @SdNiharika]

That weekly temptation to modify my email signature to

Best,

Dr. {firstname=contactfield}

in appreciation of the predatory journals that have such immense faith in my ability to conduct research.

[Jay Cummings @LongFormMath]

Here is a lovely geometric proof that sqrt(2) is irrational.

The proof begins by looking at an isosceles right triangle whose legs have length 1. By the Pythagorean theorem, this triangle's hypotenuse will have length sqrt(2).

[GRAIL workshop @ MICCAI 2026 @GRAIL2025]

📣 GRAIL 2024, the 6th international workshop on "GRaphs in biomedicAl Image anaLysis", has been accepted as MICCAI 2024 workshop!📣

We're thrilled to announce that GRAIL will be held in-person in Marrakesh, Morocco on October 6th!

#MICCAI2024 #GRAIL2024 #geometricdeeplearning

[Niharika S. D'Souza @SdNiharika]

@KaustavBera11 @AnsethGroup @KKLabJHU @CUEngineering Congratulations, @KaustavBera11 !

[Mario Figueiredo @mariotelfig]

My favorite illustration of how geometric intuition fails miserably in high dimensions.