The conventional way to do SLU is to convert the #speech into text, and then convert the text into the intent. For a great example of this type of system, see this paper by alice coucke and others: https://arxiv.org/abs/1805.10190
Another approach is end-to-end SLU, where the speech is mapped to the intent through a single neural model. End-to-end SLU: -is simpler, -maximizes the actual metric we care about (intent accuracy), -and can harness info not present in the text, like prosody (e.g. sarcasm).
End-to-end #SLU is theoretically nice, but learning to understand speech totally from scratch is really hard—you need a ton of data to get it to work. Our solution: transfer learning! First, teach the model to recognize words and phonemes; then, teach it SLU.
Some people at GoogleAI and fb research have been doing some excellent work on end-to-end SLU, but without access to their datasets, it's impossible for most people to reproduce their results or do any useful research.
So we created an SLU dataset, Fluent Speech Commands, which http://Fluent.ai is releasing for free!

It's a simple SLU task where the goal is to predict the "action", "object", and "location" for spoken commands.
We hope that you find our dataset, #PyTorch code, pre-trained models, and paper useful. Even if you don't want to do SLU, the dataset can be used as a good old #classification task, adding to the list of open-source #audio datasets. Enjoy!

✴️ @AI_Python_EN

Despite attempts at standardisation of DL libraries, there are only a few that integrate classification, segmentation, GAN's and detection. And everything is in #PyTorch :)

https://lnkd.in/eTsqKWZ

#ai #objectdetection #machinelearning #gpu #classification #dl

✴️ @AI_Python_EN

The ability to deal with imbalanced datasets is a must-have for any #datascientist. Here are 4 tutorials to learn the different techniques of handling imbalanced data:

How to handle Imbalanced #Classification Problems in #MachineLearning? - https://buff.ly/2sIsR0M

Investigation on Handling Structured & Imbalanced Datasets with #DeepLearning - https://buff.ly/2MpxuG1

This Machine Learning Project on Imbalanced Data Can Add Value to Your #DataScience #Resume - https://buff.ly/2Mpr2i0

Practical Guide to deal with Imbalanced Classification Problems in #R - https://buff.ly/2MrS8Fr

✴️ @AI_Python_EN

#Detection ...? or #Classification ...? its all there in #PyTorch.
A pytorch lib with state-of-the-art architectures, pre-trained models and real-time updated results.
#deeplearning #ai #cnn

https://lnkd.in/eTdvfEp

✴️ @AI_Python_EN

Self-Paced Learning:
- supervised method from 2010 #NIPS
- idea: start learning with the easiest samples first and only then learn the difficult ones
- distinct from curriculum learning, where samples are pre-classified to easy/hard: we need to decide the order on our own
sample in a latent model (outliers will be the hardest)
- a better measure (!): how good are the initial predictions for the sample (samples far away from the decision boundary are the easiest).

- for #classification, samples are only easy in context of other samples!
- the set of easy samples is iteratively enlarged
- results: outperforms CCCP in #DNA Motif Finding, handwritten digit recognition and others problems
- link: https://papers.nips.cc/paper/3923-self-paced-learning-for-latent-variable-models