Shazam for Wails: This App Can Tell You the Reason Why Baby Is Crying
May 11, 2018
From “I’m hungry!” to “Change my diaper!”, baby can be trying to communicate any number of things when she bursts into tears. And it can be frustrating as anything to decipher baby’s cries before she’s developed early speech skills. That universal parenting frustration is what motivated one UCLA computational neuropsychologist to make it her mission to help parents better understand what’s making baby upset–with the simplicity of an app. Yep, Ariana Anderson, PhD (and mother of four) developed a system of algorithms that decodes baby’s cries to a tee, all wrapped up in a handy app named ChatterBaby. And with a 90 percent accuracy rate for pain-related cries, she’s getting pretty darn close to reading baby’s mind.
According to ChatterBaby’s website, the app “works by analyzing the specific sounds that your baby makes when he or she is crying and comparing them with over 2,000 other sounds in [their] infant sound database.” The app then translates baby’s cries into a need that can be fulfilled. That probably sounds pretty amazing, especially for new parents who are just getting started and don’t have years of professional baby detective work under their belts.
Curious how Anderson actually built her algorithms? With the help of real moms, of course. The team at UCLA worked with “veteran mothers” to “meticulously label” types of recorded cries. Labels included fussy, hungry, separation anxiety, colic and scared. Only the cries that were universally labeled by the veteran moms as the same “type” of cry were used in the algorithm. The app currently predicts three cry types: hunger, pain and fussy. Along with including more specific labels for cries in the app, Anderson hopes to eventually use cry patterns to help diagnose autism earlier.
There are a few caveats to using the app that data-wary parents should make note of. Parents are required to sign a consent form prior to using ChatterBaby, releasing all recordings to the app’s founder Anderson and her team at UCLA. Their website explains, “anytime you predict using ChatterBaby, you are sending your audio data to UCLA along with any other data you provide (baby’s age, etc…) and are consenting to be in our research study.” Since the app is built using an artificial intelligence algorithm, it constantly requires data to make updates, and eventually make better predictions.
ChatterBaby might not be for every parent, but it could definitely come as a comfort late at night when you’re at your wits’ end. If apps just aren’t your thing, check out expert tips for what to do when baby gets fussy at night.