Training camps

Training camps 2021 - 2022

 
*** Training Camp by Leonardo and the Leonardo Labs ***
Company website: https://www.leonardo.com/en/innovation-technology/leonardo-labs
Date: September 7th-9th, 2022

Title: Image Classification with Real World Data Distributions

Abstract:
Image classification is one of the most fundamental tasks in computer vision, the objective is to predict the class of an image by assigning it to a specific label

Sometimes this task is easy, other times very hard. In real applications, building high-performance and robust models can be challenging because of the poor quality of the data in terms of:
- errors in image labels (noisy datasets)
- dataset biases
- very similar classes in appearance (fine-grained classes)
- lack of images for one or more specific classes (imbalanced datasets)

The main aim of this Training Camp will be to enable students to build their own image classifier applied to vessel classification and to participate in an online Kaggle competition organized for the course participants.

What you will learn (codes are written using Pytorch, Torchvision and scikit-learn):

- Convolutional Neural Networks for computer vision
- Exploratory data analysis and validation strategies
- Handling imbalanced datasets
- Augmentation techniques in computer vision
- Results analysis and dataset cleaning

Pre-requisites:
- Solid knowledge of Python and Jupyter Notebook
- Knowledge of neural network frameworks like Pytorch (preferred), Tensorflow, and Keras are admitted
- Basic knowledge of Python data science tools like Pandas and scikit-learn

 
The winners of the Leonardo training camp are:
 
Seyed Behdad Ahmadi
Giovanni Giunta
Giuseppe Semeria
Saeed Zohoorianmoftakharkhodaparast
Mehrzad Jafari Ranjbar

 

Congrats!

20220909_155923.JPG20220909_160733.JPG20220909_160312.JPG20220909_161252.JPG
 
 
*** Training Camp by Poste Italiane ***
Company website: https://www.poste.it/
Date: July 6th-8th, 2022

Title: Driving Business Decisions through AI

Abstract:
Algorithmic Business Thinking is a paradigm defining algorithms bases on a symbiotic cooperation of humans and machines working side-by-side in mitigating the risk of unconscious biases that could cause effectiveness loss on the final decision. The partnership of humans and machines derives AI driven business decisions in a faster and more effective way, supporting better scaling when heterogeneous business use-cases demanding increases. AI algorithms will improve marketing strategies and drive product evolutionary transformation, embedding the AI in the product itself or using AI to design for innovation.

On the other hand, in some cases this paradigm may raise the need for an ethical consilience among decisions provided by machine vs human ( "Is this the right thing to do; what are the unintended consequences?")

This camp is aimed at the accomplishment of the following targets:
- Collect the data provided during the onboarding of a customer purchasing energy supply offer in order to predict the energy consumption in the next period of time,
- Drive the identification of a business decision (i.e. commercial offer), and ensure consilience of sustainability and consumption demand.

As a prerequisite for this training, students must have solid knowledge of Python and machine learning most frequently adopted libraries.