Questions tagged [machine-learning]

The tag has no usage guidance.

Filter by
Sorted by
Tagged with
1 vote
3 answers

BirdNET: workflow for correcting manually trained classifier?

I am using BirdNET GUI for analysis of bird vocalizations. I have created a classifier using the GUI and some training data, and I am wanting to improve it, e.g. correct false classifications and ...
Emma's user avatar
  • 11
5 votes
4 answers

How to make class activation maps from audio

I am interested in class activation maps to interpret the results of classifications by a deep neural network, as shown in pages 173-180 (Figures 5.8 to 5.15) of the thesis of S. Kahl (2020) https://...
Camille Rondeau's user avatar
9 votes
4 answers

Filtering out wind from audio recordings

I am currently working on building a deep learning model for snowscooter detections. The model does fairly well at detecting snowscooters but it triggers on wind noises, increasing the rate of false ...
Benjamin Cretois's user avatar
11 votes
1 answer

Are there available raw model weights for bat detector / classifier?

I will need to analyze some soundscape data from Norway and it is likely that bats are present in the area. I was wondering if there was any available model trained for bat detection or ideally bat ...
Benjamin Cretois's user avatar
7 votes
1 answer

Data augmentation in bioacoustics: pros and cons?

Even though we already have a discussion regarding what kind of data augmentation makes sense for ML in bioacoustics (Data augmentation strategies for bioacoustics machine learning), i believe it ...
lostanlen's user avatar
  • 393
8 votes
2 answers

What are the differences in selection table data in Raven Pro versus Raven Lite when labeling vocalizations?

For the purposes of creating a library of labeled vocalizations for suite of species to use in a machine learning algorithm - is there a difference in using Raven Lite versus Raven Pro in terms of the ...
ameliaduvall's user avatar
13 votes
2 answers

Should we normalize audio before training a ML model

In computer vision it seems that image normalization is very important to keep the gradients "in check" (detailed description here). However, I was wondering if it was also the case in audio ...
Benjamin Cretois's user avatar
17 votes
3 answers

Harbour porpoise click detection using deep learning

Harbour porpoises produce distinct narrow band high frequency clicks centered around 130 kHz. Typically these are usually easy to identify by a manual analyst and should be relatively easy to ...
user avatar
8 votes
1 answer

Apple M1/M2 chips for acoustic deep learning

I’ve been reading that both Tensorflow and Pytorch deep learning frameworks are now supported by Apple M1/M2 chips. This means that a deep learning acoustic species classifier should run an order of ...
user avatar
17 votes
4 answers

Convolutional Neural Network (CNN) where to start?

I am finally getting into signal identification and would like to test Convolutional Neural Network (CNN). I am working in R and willing to do Python. What packages are you using in those two ...
Amandine Gasc's user avatar
  • 1,512
12 votes
2 answers

What criteria should be used to decide whether an automated detection overlaps with a reference annotation?

For those who use automated detection and routinely need to evaluate detector performance, what constitutes an overlap between an automated detection and a reference annotation? Most of my ...
Cathleen B's user avatar
  • 1,014
9 votes
1 answer

Probability of detection / probability of false alarm vs. precision / recall

When using automated detection of focal species sounds, when and why should we focus on characterizing probability of detection and probability of false alarm, vs. focusing on precision and recall? ...
Cathleen B's user avatar
  • 1,014
10 votes
2 answers

Why can't I run a classifier on the dataset on which it was trained?

I'm new to machine learning methods and have terabytes of acoustic data containing odontocete clicks. I want to train a classifier to classify these clicks, so I've used a subset of this data to do so....
Chloe's user avatar
  • 3,685
16 votes
2 answers

Data augmentation strategies for bioacoustics machine learning

When training a Neural Network for bioacoustics classification tasks I usually follow a data augmentation pipeline which is composed of: Changes in sound to noise ratio: for instance, if my sound ...
Benjamin Cretois's user avatar
24 votes
6 answers

Is there any library that gathers marine species vocalizations?

I would like to automatize the analysis of marine soundscapes by training a neural network that can detect and identify marine species. However, I need a large amount of training data to build such ...