In this episode I briefly explain the concept behind activation functions in deep learning. One of the most widely used activation function is the rectified linear unit (ReLU).
While there are several flavors of ReLU in the literature, in this episode I speak about a very interesting approach that keeps computational complexity low while improving performance quite consistently.
This episode is supported by pryml.io. At pryml we let companies share confidential data. Visit our website.
Don't forget to join us on discord channel to propose new episode or discuss the previous ones.
ReferencesDynamic ReLU https://arxiv.org/abs/2003.10027
Building Self Serve Business Intelligence With AI and LLMs at Zenlytic (Ep. 235)
Money, Cryptocurrencies, and AI: Exploring the Future of Finance with Chris Skinner (Ep. 234)
Debunking AGI Hype and Embracing Reality (Ep. 233)
Full steam ahead! Unraveling Forward-Forward Neural Networks (Ep. 232)
The LLM Battle Begins: Google Bard vs ChatGPT (Ep. 231)
Unleashing the Force: Blending Neural Networks and Physics for Epic Predictions (Ep. 230)
AI’s Impact on Software Engineering: Killing Old Principles? [RB] (Ep. 229)
Warning! Mathematical Mayhem Ahead: Demystifying Liquid Time-Constant Networks (Ep. 228)
Efficiently Retraining Language Models: How to Level Up Without Breaking the Bank (Ep. 227)
Revolutionize Your AI Game: How Running Large Language Models Locally Gives You an Unfair Advantage Over Big Tech Giants (Ep. 226)
Rust: A Journey to High-Performance and Confidence in Code at Amethix Technologies (Ep. 225)
The Power of Graph Neural Networks: Understanding the Future of AI - Part 2/2 (Ep.224)
The Power of Graph Neural Networks: Understanding the Future of AI - Part 1/2 (Ep.223)
Leveling Up AI: Reinforcement Learning with Human Feedback (Ep. 222)
The promise and pitfalls of GPT-4 (Ep. 221)
AI’s Impact on Software Engineering: Killing Old Principles? (Ep. 220)
Edge AI applications for military and space [RB] (Ep. 219)
Prove It Without Revealing It: Exploring the Power of Zero-Knowledge Proofs in Data Science (Ep. 218)
Deep learning vs tabular models (Ep. 217)
[RB] Online learning is better than batch, right? Wrong! (Ep. 216)
Create your
podcast in
minutes
It is Free
Insight Story: Tech Trends Unpacked
Zero-Shot
Fast Forward by Tomorrow Unlocked: Tech past, tech future
Black Wolf Feed (Chapo Premium Feed Bootleg)
Bannon`s War Room