Dmytro Tkanov

Snr Eng Mngr @ SoundHound
ASR research

[Max Ved @_madved]

“Nowadays the goal of #TTS — the #Text-to-Speech conversion #technology — is not to simply have #machines talk, but to make them sound like humans of different ages and gender.” — @_sciforce_ https://t.co/Eh7DwkH5nB

[Dmytro Tkanov @akademi4eg]

Going to give a talk today on tradeoffs in machine learning model design. #datamass #Gdansk

[Dmytro Tkanov @akademi4eg]

Poster is up and running :) #Interspeech2019

[Yann LeCun @ylecun]

ICLR h5-index = 150!
150 ICLR papers received over 150 citations over the last 5 years.
Ranked 42 among all scientific publications.
This is quite remarkable given that ICLR has only been in existence for 6 years...... https://t.co/3BVHHhdJHC

[Dmytro Tkanov @akademi4eg]

@g3n1uss У всех права на публикацию лиц спросил? :)

[Dmytro Tkanov @akademi4eg]

Our paper "Embedding-based system for the Text part of CALL v3 shared task" is accepted to SLaTE 2019 (https://t.co/LBsvWTxrBz)

[Dmytro Tkanov @akademi4eg]

Paper and code of our recent Interspeech submission: https://t.co/9gt1537s4V and https://t.co/wFISlxmpx4

[Dmytro Tkanov @akademi4eg]

Our paper "Attention model for articulatory features detection" is accepted to #interspeech2019

[Greg Stoddard @gregstod]

Movie exec: Pitch me
Me: It’s a movie about high school girls trying to figure out what clique they belong in. They move from clique to clique and eventually stop when they minimize their differences. It’s called K-Means Girls.
Movie exec: Get out of my office

[Dmytro Tkanov @akademi4eg]

It works! :) #is2019 @ AmberExpo https://t.co/1dOU6rVDrA

[Linux @Linux]

25 years ago on 14th March 1994, Linux kernel 1.0.0 was released. It consisted 176,250 lines of code.

https://t.co/6boAtZOpOA

[TensorFlow @TensorFlow]

Learn how to build AI that works offline with Coral Dev Board and TensorFlow Lite!

Read the post by @dansitu ↓ https://t.co/aTcdqhtaK0

[Dmytro Tkanov @akademi4eg]

Lyon's auditory model port from MATLAB/C to Python.
https://t.co/wJpwH388yZ

[Dmytro Tkanov @akademi4eg]

@_madved @intel Sounds like a hosted Kubeflow. Also interesting how it compares against Google CMLE.

[Dmytro Tkanov @akademi4eg]

I liked a @YouTube video https://t.co/l7DPpNdl7d Stop

[Andrew Ng @AndrewYNg]

Announcing CheXpert! Large dataset of chest X-rays co-released with MIT's MIMIC-CXR dataset. Join our competition to test your chest X-ray interpretation model: https://t.co/ldWBwjoZKn

@jeremy_irvin16 @pranavrajpurkar @mlko53 @curtlanglotz @mattlungrenMD @Stanford #AAAI19

[Eric Jang @ericjang11]

Holy Jit, https://t.co/cPYL24r00c is amazing! Stay tuned for a cool tutorial blog post that uses it :)

[Dmytro Tkanov @akademi4eg]

Got into #TFRC program! Uploading 210Gb to #GCP to test TPU training. Probably it would have been wiser to do data preparation directly in the cloud...

[Dmytro Tkanov @akademi4eg]

Don't miss out on the "https://t.co/69Wh1ieHum 2019" https://t.co/F7eQnGSuNU

[Max Ved @_madved]

“Most of the #CALL solutions use #ASR at their back-end. However, a conventional ASR system trained on native #speech is not suitable for this task, due to students’ accent, language errors, lots of incorrect words or out-of-vocabulary words…” https://t.co/ijQwYVeZ4M