Ah the week after Thanksgiving… America’s version of running with the bulls in Pamplona 🇪🇸 only here there are no bulls, just crazed discount shoppers, who will and have gorged fellow deal seekers with whatever sharp tool was within arm's length. Can someone please make a Bot for this already? Does Pepper have a Black Friday/shoving other deal-seekers setting yet?
Oh retail. Back at it again with a foray into leveraging Messenger to drive commerce, this time from Saks with their second-annual gift recommender. Few things to consider here: 1) Admittedly, we can be quite skeptical about Retail over chat--we're just not sure consumers are ready to purchase this way, however we are huge fans of bots for curation, which leads us to 2) Saks is so high-end, it naturally attracts an older audience--are they really on Messenger? 3) Considering they're bringing this back for a second year in a row--we guess the answer is yes...@Saks-share your results, will ya?!
For you NLP engineers and developers out there, the folks at Hugging Faces released a paper and the training code from over a year's worth of experimenting with Multi-Task Learning. The HMTL model outperforms contemporary NLP tasks and use cases. 👀
Is it just us? Or does anyone else get tingly when a golden use case for Bots like this gets funding? Truly, this is AI/Bots at its finest--taking repetitive, error-prone tasks and automating it. Between this and Leena AI tackling HR, we may be well on our way to botifying the backbones of all companies. Now who's cookin' up the Bot janitor…
Google released over two MILLION human-labeled 10-second sound clips drawn from You Tube videos. Over 600 audio event classes. One more step chipping away at the hardest part of AI: dirty data and labeling. Groan.
Find this interesting? Forward it to a friend. Curated with :love: from NY by Clara de Soto (@Claradactyl), Omar Pera (@ompemi) and the Reply.ai team. Ping us to chat ;)
Chat Bots Weekly · 305 West Broadway, Suite #176 · New York NY 10013 · USA