Aleksandr Kovalev (@koval_alvi) 's Twitter Profile
Aleksandr Kovalev

@koval_alvi

PhD Student, build visual neural interfaces at @NeuroEngin_UMH
Co-Founder of @alvilabs

ID: 1451507288741658630

linkhttps://www.linkedin.com/in/kovalalvi/ calendar_today22-10-2021 11:14:54

754 Tweet

734 Followers

1,1K Following

Aleksandr Kovalev (@koval_alvi) 's Twitter Profile Photo

I highly recommend participating! Last year we tried to finetune whisper, hope this year we could try smt more interesting 🤔

Mehdi Azabou (@mehdiazabou) 's Twitter Profile Photo

Excited to announce the Foundation Models for the Brain and Body workshop at #NeurIPS2025!🧠 We invite short papers or interactive demos on AI for neural, physiological or behavioral data. Submit by Aug 22 👉 brainbodyfm-workshop.github.io

Excited to announce the Foundation Models for the Brain and Body workshop at #NeurIPS2025!🧠

We invite short papers or interactive demos on AI for neural, physiological or behavioral data.

Submit by Aug 22 👉 brainbodyfm-workshop.github.io
Instituto de Bioingeniería (@ib_umh) 's Twitter Profile Photo

📢HOY, 14 de julio de 2025, cita obligada en el Instituto de Bioingeniería 🔝 ¡¡Seminario!! 🗣️" SCAPE: Shift-variant Cortical-implant Adaptive Phosphene Encoding, impartido por Michelle Appel del  Institute for Brain, Cognition, and Behaviour, Radboud University (AI Department), the Netherlands.

📢HOY, 14 de julio de 2025, cita obligada en el <a href="/IB_UMH/">Instituto de Bioingeniería</a> 🔝 ¡¡Seminario!!

🗣️" SCAPE: Shift-variant Cortical-implant Adaptive Phosphene Encoding, impartido por Michelle Appel
del  Institute for Brain, Cognition, and Behaviour, Radboud University (AI Department), the Netherlands.
Aleksandr Kovalev (@koval_alvi) 's Twitter Profile Photo

meg speech decoding 2025 You have one week to make a submission for the first track. Deadline: 31st of July We extend our conv baseline with the temporal layer. It improves our solution by 5% Now we're the second. Continue to train models. neural-processing-lab.github.io/2025-libribrai…

meg speech decoding 2025

You have one week to make a submission for the first track.
Deadline: 31st of July 

We extend our conv baseline with the temporal layer. It improves our solution by 5%

Now we're the second.

Continue to train models.

neural-processing-lab.github.io/2025-libribrai…
Aleksandr Kovalev (@koval_alvi) 's Twitter Profile Photo

MEG speech decoding update v3 The first binary classification track has ended. We made it to the top 3 on the public leaderboard. Our best model is conv + transformer encoder with rope. Almost everything I tried as extra improvements didn't work😅 I'll share details later

MEG speech decoding update v3

The first binary classification track has ended.

We made it to the top 3 on the public leaderboard. Our best model is conv + transformer encoder with rope.

Almost everything I tried as extra improvements didn't work😅

I'll share details later