top of page
The brains factory(1).png

Discover the two biggest brains in the world: the human brain and the artificial brain.


Who am I ?

Lydia Bessaï

Data scientist and Neuroscience expert


Blog Author of

The Brains Factory


About The Brains Factory

The Brains Factory is the place where all brains are celebrated (artificial brains and human brains). The Brains Factory offers popular content in neuroscience and artificial intelligence. Through articles and podcasts, you can learn more about your brain or discover several topics around artificial intelligence (data science, machine learning, etc.).
The Brains Factory also offers a service of creativity workshops. By mixing neuroscience and marketing techniques, you will be able to develop your creativity with scientific tools.


❓ 𝐂𝐨𝐦𝐦𝐞𝐧𝐭 𝐨𝐩𝐭𝐢𝐦𝐢𝐬𝐞𝐫 𝐯𝐨𝐭𝐫𝐞 𝐑𝐀𝐆 𝐞𝐧 𝐭𝐚𝐧𝐭 𝐪𝐮𝐞 𝐃𝐚𝐭𝐚 𝐒𝐜𝐢𝐞𝐧𝐭𝐢𝐬𝐭 ❓

🆘 6 conseils ici sur ce post !

💻 Pour rappel, RAG= Retrieval Augmented Generation. C'est une technique qui permet d'optimiser les résultats des LLMs pour cibler plus précisement les données qui sont en lien avec la question de l'utilisateur.

💭 Une superbe métaphore dans l'article sourcé :
"In layman terms, RAG is to LLMs what GPS is to drivers. GPS is an external data source, that drivers can use to look up relevant information about the route of the destination. "

🚮 𝐂𝐨𝐧𝐬𝐞𝐢𝐥 𝐧𝐮𝐦 1 : Travaillez sur vos données.
Eviter le "garbage in, garbage out" : trier et nettoyez vos données. L'efficacité du RAG, et donc la qualité de réponse des LLMs, dépend directement des données injectées. Par exemple, des doublons sont souvent présents dans les contexte des LLMs ce qui réduit la fenêtre de connaissance et la pertinence des données de ce dernier.

☯ 𝐂𝐨𝐧𝐬𝐞𝐢𝐥 𝐧𝐮𝐦 2 : Recherche par similarité vectorielle ou par mot-clef ?
Testez puis choisissez la meilleure stratégie d'indexation. L'indexation par similarité vectorielle n'est pas la solution par défaut. Parfois, l'indexation sur la base d'une recherche par mot-clefs est plus efficace notamment dans le cadre de recherche d'item particulier. A contrario, la recherche par similarité vectorielle est plus efficace pour les recherches globales.



🚨 What if the brain and machines reasoned the same way?

👉 Classic AI systems (like machine learning or generative models), are systems that need to train with data. After practice comes learning. But what is the difference with the brain?

👉 The human brain has two learning systems: a bottom-up system and a top-down system. The bottom-up system takes data from the outside (bottom) to bring it back to the brain (up). The brain will train thanks to this. The top-down system is a Bayesian system that makes assumptions about the world and corrects our potential predictions. For example, if I learned that birds can fly and I come across an ostrich, I will assume that bird flies. Seeing that I am wrong, I will modify my hypothesis of the birds to adapt it to reality. This is the strength and difference of the human brain vis-à-vis the machine.

Funny fact
Top 3 articles

The top 3 articles

Find our most popular articles here !



History of a fabulous encounter between artificial intelligence and neuroscience

Structures en papier


Creativity explained through cognitive science

Image d'IRM


Right or wrong: The neuromyths !

How have neuroscience and artificial intelligence developed together over time?

What does the concept of creativity mean in cognitive science? How to measure creative potential?

What are the biggest biases in neuroscience told? Let's get the truth together in this article.

Dernier podcasts

The latest podcast

PODCAST #4.png

🎙 On a scale of 1 to 10: how human would you rate ChatGPT?

🤖 How human can ChatGPT be? Is he able to put himself in someone else's shoes? Is he able to have an intuitive reasoning?

🧠 I tested ChatGPT with 4 cognitive tests to answer these questions. The result in an article and in my latest podcast.

Be careful, french only !


Creativity workshops

Offer a quality creativity workshop to your employees.



Atelier de créativité
bottom of page