Data Science Milan #004

Dear Data Science Milan Community,

Have you ever wondered if you missed out on the groundbreaking transformer (and lately Generative AI) trend back in 2017 when Google introduced the Transformer architecture in that paper? If so, don't worry, we've got you covered. Let's take a quick journey back to that pivotal moment in the field of data science:

• Changing of the Guard: In 2017, Natural Language Processing (NLP) was dominated by LSTM networks that were pretty performing in language translation tasks (spoiler they still are). Anyway, transformer architecture marked a significant shift by replacing these LSTMs with an Encoder-Decoder model, leading to a substantial improvement in NLP tasks.
• The Attention Mechanism: The Transformer's key innovation was the attention mechanism. This mechanism enables the model to focus on relevant parts of the input sequence, rather than processing it sequentially. This breakthrough significantly improved both training efficiency and model performance.
• Transfer Learning: The paper also introduced the concept of transfer learning in NLP tasks (mostly borrowed from computer vision). Pre-training the model on vast amounts of data and fine-tuning it for specific tasks became a game-changer, as it allowed for more efficient and accurate NLP models.

Excitingly, in our next newsletter, we'll start providing practical examples of how Transformer architectures are revolutionizing common NLP tasks. Stay tuned for insights into the cutting-edge applications of Generative AI in our data science community. Don't miss out this time!

Data?Science?Milan?events

Upcoming:?On November 7th, 2023, Andrea Maiorana will speak about how data science works at Bending Spoons.?

Bending Spoons is a leading tech company based in Italy which is specialized in software and app development. Andrea will go through data science workflows at Bending Spoons, and then dive into the measurement and predictions related to app users metrics. There will be also a poster session during networking aperitivo.

Enroll to the event

Building GenAI applications with LangChain on GCP

On June 27th, 2023, Ivan Nardini from Google Italia, Ivan Vigorito and Domenico Vitarella from Generali showed us an interesting Generative AI use case. ?

With an introduction from Ivan about LangChain, a framework that simplifies 'prompt piping' with large language models (LLMs) using Google Cloud. Then Ivan and Domenico shared a case study of using LangChain with PALM 2 for knowledge base Q&A in Generali. The magic is done by giving PALM 2 the ability to query a 'chunkenized' version of different documents using LangChain to 'connect' the information that is not available together.

TinyML

On June 15th, 2023, Pietro Montino spoke about TinyML, a core technology of modern IoT systems where data are analyzed close to where they are collected,?with benefits for energy efficiency, latency, privacy, and cost.

Knowledge section

Here are selected resources about app development:

• Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., ... & Polosukhin, I. (2017). Attention is all you need. Advances in neural information processing systems, 30. - https://doi.org/10.48550/arXiv.1706.03762
• TUNSTALL, Lewis; VON WERRA, Leandro; WOLF, Thomas. Natural language processing with transformers. " O'Reilly Media, Inc.", 2022.

Be involved!

We want also to remind you that if you like and enjoy our events, you can get in touch with us at?[email protected]?to be involved in organizing new great online activities.

We are also very happy if you are interested in being a speaker or if you want to share your expertise or experience with the?Data?Science?Milan?community!!!

Wallboard

Would you like to become one of our sponsors and increase your popularity among the?Data?Science?community? Write?here

If instead, you would like to promote a message to the wallboard, please contact us and send us your relevant announcements. We will publish them here.