Pratinav Seth
@pratinavseth
Followers
815
Following
9K
Media
18
Statuses
213
AI explorer Interpretability, Alignment, Optimization, Safety & More at AryaXAI | AI for Social Good | AAAI UC 23 Scholar | Prev. @ Mila,Bosch,Manipal.
Kolkata, India
Joined July 2020
In this work we show how interpretability methods can guide pruning, making models more efficient while preserving accuracy & transparency. Authors: Nikita Malik, Pratinav Seth, Neeraj Kumar Singh, Chintan Chitroda, Vinay Kumar Sankarapu . Pre-Print :
arxiv.org
Deep learning has driven significant advances in medical image analysis, yet its adoption in clinical practice remains constrained by the large size and lack of transparency in modern models....
0
1
3
Hi Everyone, I will be presenting our work "Interpretability-Aware Pruning for Efficient Medical Image Analysis", today at the 1st MICCAI Workshop on Efficient Medical AI! #MICCAI2025 #MedicalAI #EfficientAI #Interpretability
1
0
16
Sadly, I will be missing #ACL2025 in Vienna 🇦🇹 for our presentation due to visa issues :(( Feel free to reach out with any questions, follow-ups, or collaborations! @pratinavseth @Adignite60 @Siddharth4848 @TanujT2005 @i_amanchadha @VinijaJain
🎉 Thrilled to share that 2 papers (1 first-author) accepted to #ACL2025 #ACL2025NLP @aclmeeting! Huge thanks to my amazing collaborators! 🙌 Preprints out soon. See you in Vienna! 🇦🇹
0
1
19
Postdoc opportunity in collaboration with our lab on downscaling/superresolution for biodiversity, details here: https://t.co/Eff3OYCEIR
community.climatechange.ai
As part of an IVADO initiative on AI and environment, a coalition of Montreal labs is recruiting a postdoctoral researcher for an interdisciplinary project applying state-of-the-art machine learning...
0
8
20
Excited to be presenting our paper at ICML 2025! 📍 West Exhibition Hall B2–B3 (W‑402) 🗓️ Wednesday July 16 | 11:00 am – 1:30 pm I'll be around through Thursday — feel free to say hi! Happy to chat about AI for social good, Interpretability, Safety, and Tabular FM. #ICML2025
I’m happy to share that our paper on detecting oil & gas wells with AI has been accepted to ICML 2025! Millions of such wells exist, leaking methane & other toxins. We show how remote sensing can help find them, using high-resolution, multi-spectral satellite imagery. @icmlconf
1
1
33
Last few days left : Call for Paper: 2nd edition of CV4DC: Computer Vision for Developing Countries Workshop at #ICCV2025. Share your innovative research on computer vision addressing challenges in developing region 📅 Deadline: June 27, 2025 📷 https://t.co/E8sQrA2keq
0
4
5
Sincere appreciation to my co-authors: @XMichelleLinX , Brefo Dwamena Yaw, Jade Boutot, @MaryKangS , @david_rolnick . #MachineLearning #ComputerVision #ClimateAction #RemoteSensing #aiforgood Look for our work at #ICML2025! @Mila_Quebec
0
0
6
Our work includes evaluating baseline AI algorithms. While results show promise for automated monitoring, they also highlight the task's complexity, inviting further research to enhance scalability & accuracy. Dataset: https://t.co/T0kKnrouPn Preprint: https://t.co/GAn8YVV2Kn
1
0
5
It is aimed to empower environmental action using remote sensing by enabling the scalable identification of such wells, assisting experts to: (1) Better estimate methane emissions and groundwater contamination. (2) Prioritize plugging and remediation of wells.
1
0
2
To address this, we developed the Alberta Wells Dataset (AWD): the first large-scale benchmark for this critical challenge. AWD provides over 213,000 well instances (abandoned, suspended, active) in Alberta, Canada, with high-resolution, multi-spectral satellite imagery.
1
0
4
The undocumented nature of countless such wells creates profound environmental uncertainty. In Canada, for example, abandoned wells are the biggest source of uncertainty in methane emissions estimates.
1
0
2
I’m happy to share that our paper on detecting oil & gas wells with AI has been accepted to ICML 2025! Millions of such wells exist, leaking methane & other toxins. We show how remote sensing can help find them, using high-resolution, multi-spectral satellite imagery. @icmlconf
1
4
54
Happy to Share Acceptance of SELF Percept at ACL 2025 🎉! Pre-Print out soon .
🎉 Thrilled to share that 2 papers (1 first-author) accepted to #ACL2025 #ACL2025NLP @aclmeeting! Huge thanks to my amazing collaborators! 🙌 Preprints out soon. See you in Vienna! 🇦🇹
2
0
33
All thanks to you @ignoreandfly 🙌! It was great working with you .
The first paper acceptance of 2025 is here. After so many iterations with @pratinavseth, the first to see the light of day has also seen the light of acceptance.
1
0
8
Was about to post a recap of my time at ICLR… but in the meantime — 1/2 papers accepted at @icmlconf 😇! A few more workshop & venue acceptances rolled in recently too — more updates in the coming days! Super grateful to my collaborators & mentors for all the support!
5
1
49
More than a day passed no response from @IndiGo6E 🙂even after sharing details. Classic Indigo Experience I assume.
Can’t believe this happened on a flight. On my recent trip from Singapore to Kolkata, my entire laptop bag got soaked(my bottle was empty). Tried reaching out to @IndiGo6E staff for help but got no response — just got brushed off with we aren’t responsible.
0
0
7
Can’t believe this happened on a flight. On my recent trip from Singapore to Kolkata, my entire laptop bag got soaked(my bottle was empty). Tried reaching out to @IndiGo6E staff for help but got no response — just got brushed off with we aren’t responsible.
3
0
18
I will be presenting my work at the Climate Change AI Workshop. Feel free to drop by and learn more .
Our #ICLR2025 workshop is about to kick off in Hall 4 #3!🎉 📺Join our FREE livestream: https://t.co/8h3W4ipZvK
0
0
14
I'll be at #ICLR2025 next week and will be presenting our paper at the CCAI Workshop (Details Soon). Lately, I have been working on AI Explainability, Alignment and Safety. I am also interested in AI for Social Good applications. If you happen to be around, would love to meet.
0
0
28
@abeirami @icmlconf In hindsight, this checkbox thing is turning out to be a bad idea. Reviewers are just clicking on the checkbox without any feedback on how the concerns remain unaddressed by the rebuttal. @icmlconf can still salvage the situation; ask ACs to mandatorily chase such reviewers.
2
4
23