The purpose of this presentation is to detail the auditory transform for AI/ML processing. The talk will begin with conventional Fourier Transform and how to use it for generating Mel-Frequency Cepstral Coefficients (MFCC) for speech recognition. We will then talk about the wavelets and wavelet transform and why wavelets are better auditory transform for non-stationary signals. The 2nd part of the talk we will be Scattering Wavelet Transform where we will detail how scattering wavelet transform could be used as an alternate to Convolutional Neural Network (CNN) for IoT devices. The material is already covered as part of my enclosed talk
Co-sponsored by: CH01072 – New Jersey Coast Section Chapter,COM19
Agenda:
Arrive 06:00PM : Test audio and video
Set up on-line for hybrid
06:25PM Introduce Speaker
06: 30PM: Talk begins
7:45PM Q & A session
8:00PM Concluding remarks and session ends
Ambiq SoC AI Audio for Wearables
Thursday, December 28, 2023
6:00 PM | (UTC-05:00) Eastern Time (US & Canada) | 2 hrs
(https://ieeemeetings.webex.com/ieeemeetings/j.php?MTID=m24f30efab90695da8e740c27ef4c07b9)
https://ieeemeetings.webex.com/ieeemeetings/j.php?MTID=m24f30efab90695da8e740c27ef4c07b9
Meeting number: 2531 036 2016
Meeting password: CKsaEVy5M37
Join from a video system or application
Dial 25310362016@ieeemeetings.webex.com
You can also dial 173.243.2.68 and enter your meeting number.
To dial from an IEEE Video Conference System: *1 2531 036 2016
Tap to join from a mobile device (attendees only)
(tel:%2B1-415-655-0002,,*01*25310362016%23%23*01*) United States Toll
(tel:1-855-282-6330,,*01*25310362016%23%23*01*) United States Toll Free
Holmdel, New Jersey, United States, 07733, Virtual: https://events.vtools.ieee.org/m/392906