Hi! Salve! Namaste! Γεια σου! 你好 ! مرحبا

LUCIAN GRUIA

Neural Networks behind AI (Ciklum GenAI Upskill Program, 2024)

Info

  • Project type Public Speaking, Tech Presentation
  • Date 17th of April 2024
  • My Role Speaker
  • Registered 81 participants
  • Topics AI, Neural Networks, Learning
  • Keywords Neuron, Synapse, Learning, Emergence, Weights, Activation function, Loss function, Gradient, Learning rate, Neural network parameters, Training, Testing, Epochs, Neural network architecture, Neural layers, Deep Learning
  • Skills developed Presentation, Speaking
  • Event recording (Password protected)

Description

What are the key factors that make a workplace appealing? Of course, there’s work-life balance, benefits, friendly colleagues, exciting projects, challenging tasks, and so on.

But what about education? Staying educated is essential—it keeps us competitive, regardless of industry, job level, or community affiliation. In an industry like software development, where technology evolves rapidly, staying informed is critical. And with the recent technological leaps, the pace is even faster.

After years of technology disrupting most industries, the software industry is now undergoing its own transformation. Generative AI is disrupting society and will likely change the world as we know it. Technological leaps are happening one after another, and this isn’t just marketing hype—it’s real.

We are living in fascinating times, and I am truly excited about it.

At the heart of this revolution are Neural Networks, specifically the Large Language Models (LLMs) that are transforming our lives today.

I feel privileged to have the opportunity to talk about this topic within an enthusiastic community. The Ciklum GenAI Upskill Program is an excellent platform for collaboration and collective learning, ensuring we stay at the forefront of technological advancements.

During this session, we spent nearly 2 hours exploring the core concepts of Neural Networks and discussed multiple perspectives—from technical aspects to sales and delivery management. It’s crucial for us to cultivate a learning culture and present a unified approach to technology when interacting with our clients.

I’m grateful to be surrounded by colleagues who are equally passionate about these topics and dedicated to staying informed.

I believe that this collaborative learning approach is one of the most effective ways to foster modern education for professionals.


🔥 Why is Understanding Neural Networks important:
Neural networks stand as one of the most powerful global drivers of impact. They provoke revolutions in entire industries alongside advancing Natural Language Understanding.
Foundational to Deep Learning, Neural Networks enable unprecedented data processing sophistication & transform Artificial Intelligence by mimicking human brain structures.
Beyond their vast applications, there's a deeper discussion. While we might rely on AI for many tasks, the question remains: who builds the AI? How can we design them correctly to bring about significant foundational changes?
Understanding neural networks isn't merely academic — it implies understanding the very essence of our digital future.
🧠 We will talk about
The session will focus deeply into the exploration of the underlying mechanics that enable Neural Networks to perform seemingly magical tasks. We'll examine the learning processes of Neural Networks and discuss the specific architectures that are best suited for Generative AI, including Transformers, Diffusion Models, GANs and others.

🎯Agenda:

  • Learning about what a Neuron is and how it functions.
  • The fundamental principles behind Neural Networks and their role in constructing Artificial Intelligence Agents.
  • Constructing a Neural Network programmatically.
  • Illustrating the emergence of the learning process.
  • The differences between Learning and Data Storage.
  • Exploring concepts like Knowledge, Understanding, and Reasoning.
  • Discussing various Neural Networks Architectures.

The aim of this session was to demonstrate how to create a neural network from scratch using Java and Python. I also used Javascript to build illustrations and animations to highlight the "magic" behind machine learning.

Moreover, the session did not have the focus on the technical aspects, but rather it was focused on sales/delivery managers profile. As counter-intuitive as it may look, I believe the approach of presenting to non technical people, tech concepts in plain english and demonstrating them, is a highly effective way to increase their understanding of fundamentals  and so to be more productive in the discussions with our partners.

This event posed several new challenges for me:

  • Presenting to a mix audience of tech and non-tech participants
  • Engaging a diverse group with varying levels of expertise and skill sets
  • Being broadcasted internally and recorded without the option for retakes

This was an experience that required intensive preparation and pushed me to overcome my limitations. I learned a lot and received valuable feedback.


References

Tech stack

  • Programming Languages: Java 17, Python 3.11, Javascript
  • Java stack: Spring Boot, Spring Web, Thymeleaf, Lombok, org.json, Log4j
  • Python stack: numpy, tensorflow, matplotlib, scikit-image
  • Javascript stack: d3js

Suggested posts