Description

Talk to microchips, i.e. end-to-end automatic spoken language understanding (SLU) for microcontrollers and microchips, while advocating for climate-friendly solutions through low-power consumption AI models.The talk would be geared towards data scientists who are interested more in learning about how audio is processed, and how models are trained with it. They will specifically gain insights on challenges faced when working with audio. From data pre-processing, to model architectures and evaluation metrics.The implications of this research extend to businesses and researchers interested in enhancing human-machine interaction and accessibility. Industries such as consumer electronics, healthcare, automotive, and education stand to benefit from our SLU solution's transformative capabilities. Furthermore, our commitment to sustainability is demonstrated through reduced training time, low power consumption, and efficient resource utilization.

Details

July 12, 2024

3:40 pm

-

4:15 pm

Bread Financial Room AB (Union AB)

Add to Calendar

Track:

AI & ML

Level:

Intro

Tags

Models
Models
AI Integration
AI Integration

Presenters

Danishta Saiyed
NLP Engineer
Think-A-Move Ltd.

Bio

Danishta Saiyed, an NLP Engineer specializing in developing high-accuracy language models for real-world applications, brings extensive expertise in leveraging advanced techniques such as text-to-speech engines, voice cloning, and proprietary noise-injection methods. Passionate about enhancing model performance in microchip environments, Danishta is eager to share her findings with the audience.

At DataConnect, Danishta will provide insights into developing and optimizing spoken language understanding models for microchips. Topics will include the utilization of quantized TensorFlow Lite models, experiment tracking via JSON files, and automated dataset pre-processing techniques. Attendees can expect to gain valuable insights into deploying models in resource-constrained environments, as well as practical strategies for improving accuracy and efficiency. Additionally, Danishta will touch upon environmentally friendly approaches, ensuring a holistic understanding of model deployment in sustainable contexts.

Don't miss this opportunity to learn from Danishta's expertise and discover innovative approaches to enhance speech recognition and understanding capabilities in microchip-based systems. Join Danishta at DataConnect for an enlightening session that promises to inspire and inform.