Leveraging Extremely Large Datasets in DMS: A Self-supervised Approach

You must be logged in to watch this session

First Name*

Last Name*

Job Title

Company Name

Country of Residence*

Work E-mail Address*

Password*

Confirm Password*

Your personal data will be used to support your experience throughout this website, to manage access to your account, and for other purposes described in our privacy policy.

Sense Media, on behalf of AutoSens, needs the contact information you provide to us to update you with information about AutoSens and our products. You may unsubscribe from these communications at anytime. For information on how to unsubscribe, as well as our privacy practices and commitment to protecting your privacy, check out our privacy policy.

Only fill in if you are not human

Leveraging Extremely Large Datasets in DMS: A Self-supervised Approach

Event: InCabin USA

| Session date: Wednesday 22nd May

Session date: Wednesday 22nd May

, 2024

Hear from:

Krishna Govindarao

Co-founder and Head of Product,

LightMetrics Inc.

Krishna Govindarao

Co-founder and Head of Product,

LightMetrics Inc.

Taking a leaf from the recent advances in LLMs, it is important to have a pretext task to build a base model that can train unsupervised or with minimal supervision on large datasets. On top of this base model, task specific DMS models can be trained on significantly smaller datasets that have high quality labels. The typical unsupervised approach is to train an autoencoder neural network (or similar). While this works well and results in better performance than not having this step, in this work, we propose a better unsupervised training algorithm to train the base model neural network for DMS tasks. In addition to showing an improved performance of this model compared to using the vanilla autoencoder approach and the “train from scratch” approach, we also show the new method results in more interpretable models.