Arsen Sheverdin

PhD student at @Helmholtz_AI and @TU_Muenchen. Interested in deep probabilistic ML. Previously AI MSc at the @UvA_Amsterdam, Physics/Maths BSc

[Arsen Sheverdin @ArsenSheverdin]

What if the points we randomly sample on the Poincaré disk of hyperbolic space are *already* word representations? Find more in our work “From Hyperbolic Geometry Back to Word Embeddings” at @aclmeeting #RepL4NLP2022 #ACL2022 workshop. Paper: https://t.co/xbptDs2EaC. [1/18]

[Arsen Sheverdin @ArsenSheverdin]

@liyzhen2 Sounds exciting! Could you please share the name of the workshop?

[Jeremias Knoblauch @LauchLab]

📢 Post-Bayesian online seminar series coming!📢

To stay posted, sign up at
https://t.co/Tx3ecXKdC8

We'll discuss cutting-edge methods for posteriors that no longer rely on Bayes Theorem.
(e.g., PAC-Bayes, generalised Bayes, Martingale posteriors, ...)

Pls circulate widely!

[Arsen Sheverdin @ArsenSheverdin]

If you wanna chat about research, DM me

[Arsen Sheverdin @ArsenSheverdin]

At ICML btw 😉

[Arsen Sheverdin @ArsenSheverdin]

@wgrathwohl Happy to chat over a coffee or a beer

[Vincent Fortuin @vincefort.bsky.social @vincefort]

Great news for all Bayesians: AABI is back this year, co-located with ICML in Vienna! Our CfP is out and the paper submission deadline is March 29th: https://t.co/6cyfwfCWRD
This year, we'll have a non-archival track for short papers and a PMLR proceedings track for long ones!

[Jeremias Knoblauch @LauchLab]

If you are considering a PhD in statistical methods & machine learning, consider joining our group at UCL---myself, @ilijabogunovic and @fx_briol have two projects linked below & there are other funding possibilities if you'd like to work on something else :)

[ICLR @iclr_conf]

@_vaishnavh We strive to vibe.

[Arsen Sheverdin @ArsenSheverdin]

@LauchLab @Ellis_Amsterdam Thank you, Jeremias!

[Arsen Sheverdin @ArsenSheverdin]

Delighted to be fortunate enough to have this opportunity! Excited about conducting research as an @Ellis_Amsterdam honours student!

[ELLIS Amsterdam @Ellis_Amsterdam]

Two MSc AI students join #ELLISAmsterdam as Honours Students https://t.co/D2GgtxIBJT @bagad_piyush @ArsenSheverdin

[Neil Band @neilbband]

Public release of the #̶N̶e̶u̶r̶I̶P̶S̶2̶0̶2̶2̶ #NeurIPS2021 (😅) RETINA Benchmark:
A suite of tasks evaluating the reliability of uncertainty quantification methods like Deep Ensembles, MC Dropout, VI, and more.

📜 https://t.co/6bnxXnaSCl
💻 https://t.co/D0jxQtmjx3
🧵below [0/N]

[Ramin @ramin_m_h]

In a new article published today in Nature MI @NatMachIntell we solved a differential equation that describes the interaction of neurons and synapses! This equation had no known closed-form solution since the year 1907!

This solution is important! 1/n

https://t.co/53qU76Ewrq

[Arsen Sheverdin @ArsenSheverdin]

@markus_with_k https://t.co/CEHSGBJUBf

[Arsen Sheverdin @ArsenSheverdin]

Awesome and so convenient! I was looking forward to a compilation like this one!

[Sarath Chandar @apsarathchandar]

This is a great list of 64 interesting papers focusing on lifelong learning, meta learning, few shot learning, transfer learning, reinforcement learning! Please do check the papers and their video presentations! @CoLLAs_Conf

[Massimo Caccia @MassCaccia]

🚨Preprint🚨

✅Interested in continual learning
✅motivated by RL agents continually learning in the real world but
✅don't want to mess with RL nor have the compute to run realistic/challenging environments

We've got you covered 😎
and we have some great/surprising news!

🧵

[Frank Nielsen @FrnkNlsn]

Hyperbolic embeddings of graphs popular in #MachineLearning because of low-distortion embeddings+convenience for downstream tasks
Statistical modeling in hyperbolic space: A "good" exponential family in the Poincaré half upper plane. #InformationGeometry
https://t.co/evsA8qpKx1

[Arsen Sheverdin @ArsenSheverdin]

Check out our poster at the @aclmeeting #RepL4NLP workshop poster session, which starts virtually on https://t.co/o6qUp4qxDI in 5 minutes!

[PyTorch @PyTorch]

Today we’d like to highlight features from functorch, a beta PyTorch library that provides JAX-inspired function transformations like vmap. (https://t.co/wShZQ74fUz)

If you’re not sure what sort of cool new things vmap allows you to do, read on to learn more!
(1/n)

[Arsen Sheverdin @ArsenSheverdin]

Tags: #ACL2022 #acl2022nlp #NLProc #NLP [18/18]

[Arsen Sheverdin @ArsenSheverdin]

What if the points we randomly sample on the Poincaré disk of hyperbolic space are *already* word representations? Find more in our work “From Hyperbolic Geometry Back to Word Embeddings” at @aclmeeting #RepL4NLP2022 #ACL2022 workshop. Paper: https://t.co/xbptDs2EaC. [1/18]

[Arsen Sheverdin @ArsenSheverdin]

This work is completed in collaboration with Sultan Nurmukhamedov, Thomas Mach, and @ZhenisbekA.
For more details, please consider reading the paper: https://t.co/Oqf5X7x03T, and visiting our poster presentation at the workshop [17/18].