Size: a a a

PiterPy Conf Сhat

2020 August 04

DK

Dmitriy Krasilnikov in PiterPy Conf Сhat
Alex Hultnér
No worries, I use Pragmata Pro, and have also used Isoveka quite a bit as my fonts for ligatures.
Thank you! I'll try those two then
источник

AH

Alex Hultnér in PiterPy Conf Сhat
I did never quite get ligatures to work well with Urxvt, so it probably matters more which terminal you use
источник

IS

Irina Saribekova in PiterPy Conf Сhat
Сейчас в эфире @andreamalhera
С темой: Applied machine learning: leak detection in water pipelines

Во время доклада отмечайте спикера и задавайте вопросы в этом чате на английском, он вам ответит.

Вопросы на русском также можно писать, их модератор задаст в прямом эфире после доклада.

Если хотите задать вопрос лично в эфире, проверяйте свою камеру и гарнитуру и пишите мне в личку :)
источник

DK

Dmitriy Krasilnikov in PiterPy Conf Сhat
Alex Hultnér
I did never quite get ligatures to work well with Urxvt, so it probably matters more which terminal you use
Yeah, I know that. I use kitty terminal on GNU/Linux, it supports ligatures well, but just not in neovim for some reason
источник

IS

Irina Saribekova in PiterPy Conf Сhat
Слушаем доклад, готовим вопросы :)
источник

IP

Ivan Petrov in PiterPy Conf Сhat
Have you cleaned outliers from raw audio sources? Was this a problem for your case?
источник

IP

Ivan Petrov in PiterPy Conf Сhat
Have you experienced any problems because of class imbalance for leakage?
источник

YV

Yuliya Volkova in PiterPy Conf Сhat
I think this question for @andreamalhera ^^^
источник

AM

Andrea Maldonado in PiterPy Conf Сhat
Irina Saribekova
Сейчас в эфире @andreamalhera
С темой: Applied machine learning: leak detection in water pipelines

Во время доклада отмечайте спикера и задавайте вопросы в этом чате на английском, он вам ответит.

Вопросы на русском также можно писать, их модератор задаст в прямом эфире после доклада.

Если хотите задать вопрос лично в эфире, проверяйте свою камеру и гарнитуру и пишите мне в личку :)
Hello everyone! Excited for your questions. Shoot! 📸
источник

AM

Andrea Maldonado in PiterPy Conf Сhat
Ivan Petrov
Have you cleaned outliers from raw audio sources? Was this a problem for your case?
Great question, thanks! So we did not do this as part of the preprocessing. But since the mics have other noises all the time. The difference between an outlier  and a leak outlier is still hearable which is what the autoencoder catches
источник

AM

Andrea Maldonado in PiterPy Conf Сhat
Ivan Petrov
Have you experienced any problems because of class imbalance for leakage?
Yes, our optimistic results reflect exactly that imbalance.
источник

AM

Andrea Maldonado in PiterPy Conf Сhat
Ivan Petrov
Have you experienced any problems because of class imbalance for leakage?
More about this coming on the evaluation part of the talk  #TryingMyHardestNotToSpoil :D
источник

AM

Andrea Maldonado in PiterPy Conf Сhat
Water leaks are infrequent and the kind of anomaly you don't want to provoke just for the sake of data balance. ^^ I am currently working on an approach for data augmentation or data generation using stochastic features for this and similar cases.
источник

AM

Andrea Maldonado in PiterPy Conf Сhat
Please keep the questions coming! I am impressed by the nice and engaged community ^^
источник

C

Chatbro in PiterPy Conf Сhat
Doctor Strange
Hi Andrea, how vulnerable the model to adversarial examples (if it is at all)?
источник

C

Chatbro in PiterPy Conf Сhat
Mikhail Abrash
why did you use decoder to count error later aka "encode -> decode -> error -> classify" and not just "encode -> features from melspect -> classification"
источник

IS

Irina Saribekova in PiterPy Conf Сhat
источник

IS

Irina Saribekova in PiterPy Conf Сhat
источник

IS

Irina Saribekova in PiterPy Conf Сhat
Ссылки также можно найти у нас на сайте https://piterpy.com/
источник

C

Chatbro in PiterPy Conf Сhat
источник