Egor Lakomkin

Research scientist at @gradiumAI, grpo'ing for voice AI
ex @Meta Superintelligence Labs

[Gradium @GradiumAI]

A new Gradium TTS model is in public beta. It reads phone numbers, emails, IBANs and time expressions correctly, natively.

Test it via the API, send feedback on your production cases, get 1M credits.

Prompt for your coding agent to start in seconds: https://t.co/gAK4fmO7uJ

[Gradium @GradiumAI]

Gradbot is Gradium's prototyping framework for building voice agents.

This week, @BhosalePratim sat down with cofounders @neilzegh and @lmazare to walk through the architecture and what matters when picking a harness.

https://t.co/IqI63dMlfD

[Laurent Mazare @lmazare]

Amazing work lead by @BhosalePratim, vibe code your own voice agent in minutes and benefit from asynchronous function calling: your LLM can trigger some tool calls while the bot is still talking with you! 🤖 🚀

[Mohammad Ali Zamani @ma_zamani]

Happy that our paper "Incorporating End-to-End Speech Recognition Models for Sentiment Analysis" with @EgorLakomkin got accepted to #ICRA2019
https://t.co/gf4JvjsFb2

[Egor Lakomkin @EgorLakomkin]

The code for our #emnlp2018 paper for collecting&filtering automatically speech samples from YouTube is here:https://t.co/hiJujgAoBy Few improvements are coming! paper: https://t.co/FlhJ7T6hYQ

[Egor Lakomkin @EgorLakomkin]

Our paper "Automatic Dataset Construction for Speech Recognition from YouTube Videos" got accepted @emnlp2018! Constructive and helpful reviews - now working on improving the paper for the final! #emnlp2018

[Egor Lakomkin @EgorLakomkin]

Yay! Our paper "On the Robustness of Speech Emotion Recognition for Human-Robot Interaction with Deep Neural Networks" got accepted at @iros_2018 @ma_zamani #iros2018 https://t.co/fl55RjjsNI

[Yonatan Belinkov @boknilev]

@EgorLakomkin on reusing neural speech representations for emotion recognition at #ijcnlp2017taipei @IJCNLP2017

[ODS.ai @ods_ai]

Congrats to our data scientists who occupied 8 out of top 10 places in Pathology Segmentation *TCO17* @topcoder #DeepLearning competition!

[Sauro @sauro]

OMG IT FINALLY HAPPENED! THE QUICK BROWN FOX JUMPS OVER THE LAZY DOG

[@fastml_extra @fastml_extra]

Using MNIST in machine learning is like playing "Stairway to heaven" in a guitar shop. Please try something else.

[yan @bcrypt]

https://t.co/DcHuadVFGg

[Mathieu Blondel @mblondel_ml]

I think people who fear AI don't know that what we call AI is usually nothing more than minimizing a loss function...

[@fastml_extra @fastml_extra]

Remember, kids: if you're doing math, keep it discrete.

RT @octonion: 5 Arrested in Math Lab Bust.

[thelousylinguist @lousylinguist]

Worth a read for anyone interested in computing meaning: From Frequency to Meaning: Vector Space Models of Semantics http://t.co/7zB0vn3ivO

[Reza Zadeh @Reza_Zadeh]

Challenge: Donate to some charity every time someone asks: "Is this better than Deep Learning?" #NIPS2014 #KDD2014

[St. Vincent @st_vincent]

If you refrain from reclining your seat in coach on an airplane, you're basically an American hero. #idontwanttosaywhatyouareifyoudont

[Clancy @occupanther]

Stanford Course in Natural Language Processing http://t.co/trK0p42wg3

#ArtificialIntelligence #creativecommons

[Edward Frenkel @edfrenkel]

Excerpt from LOVE AND MATH about using math in decision-making (a project I did with a doctor as a student in Moscow) http://t.co/oC4vP17r2c