Loading Events

« All Events

  • This event has passed.

DSPT#67 – Recommended By Data Scientists: Come Check it Out! (Lisboa)

December 3 @ 18:30 - 20:30

Detalhes

Getting close to Christmas and fed up of the checkout lines and seeing your budget run out? We recommend an evening with DSPT where you’ll be able to hear about how data science is helping companies develop self checkout systems from Filipe Ferreira while Kelwin, a DSPT veteran, will speak to us about how big companies leverage recommendations systems beyond the simple formats. Come cheer the end of the data science year with us!

=== SCHEDULE ===

The preliminary agenda for the meetup is the following:
• 18:30-19:00: Welcome and get together
• 19:00-19:30: Talk 1: “Low-budget Self-Checkout Development: The use of synthetic datasets to create highly scalable perception models in the wild” by Filipe Ferreira
• 19:40-19:45: Group photo
• 19:45-20:15: Networking / Coffee Break
• 20:15-20:45: Talk 2: “Providing relevant recommendations beyond the explored frontiers” by Kelwin Fernandes
• 20:50: Closing, hanging out and some beers
• 21:00: Dinner is optional but it might be an excellent opportunity for networking (https://doodle.com/poll/5zf3779bzdzx7ews)

This meetup is sponsored by NOS (https://www.nos.pt/)
See you there!

Talk 1: Low-budget Self-Checkout Development: The use of synthetic datasets to create highly scalable perception models in the wild

Abstract: In this talk we’ll show how synthetic datasets can be used to build highly scalable robust perception models in real applications. We’ll show how part of our solution for retail seamless checkout is based on models trained with synthetic data rendered from a realistic (or not) model of our environment. We’ll also show how this approach leverages a smaller size of the models and can easily scale for different environments.

Short Bio: Filipe has a master’s degree in Electrical Engineering from FEUP, where he first interacted with Computer Vision methods and applications. In his master thesis, he used Indoor Soccer footages from a Drone to analyze individual and collective behavior during matches. Afterwards, he had the opportunity to apply his knowledge in two real-life problems: first, in Follow Inspiration he helped develop the perception system that allowed a robot to follow customers in the shop so it could carry their purchases; in HealthyRoad, Filipe applied Deep Learning techniques to increase the accuracy of their driver drowsiness detection system. After a short stay in BioImgLab@INESC where he developed a 3D Convolutional Neural Network model for Lung Lobe Segmentation in Thoracic CTs, he joined Sensei as a Computer Vision/Machine Learning Engineer where he’s developing a technology that allows retailers to understand what customers do in their stores.

Talk 2: Providing relevant recommendations beyond the explored frontiers

Abstract: With new products and technologies being constantly developed, how can we sail across the sea of the unknown?
In a subscription company such as NOS, the endeavor of bringing the offer that better suits the client’s needs is a challenging one. Here, you need to make each interaction with the client meaningful, while recommending products you just released. In this talk, we will discuss how we built recommender systems that embrace cold start as an unavoidable reality.

Short Bio: Kelwin Fernandes holds a Ph.D. in Computer Science from Universidade do Porto (2018) and a Computer Engineering degree from Universidad Simón Bolívar (2012). He is the CEO of NILG.AI, a consulting company in Data Science working in multiple industries, from healthcare to telecommunications. He has been working together with NOS to scale up Data Science practice, with impactful projects ranging from ML and RecSys to NLP.

Details

Date:
December 3
Time:
18:30 - 20:30
Event Categories:
, ,
Event Tags:
, , ,
Event Website

Organizer

DSPT
Email:
info@datascienceportugal.com
Organizer Website

Venue

NOS – Edificio do Campo Grande
Rua Actor António Silva, 9
Lisbon, 1600-428 Portugal
+ Google Map