Les 20 questions du quiz
Question 1 : Which programming language was created by Guido van Rossum?
Réponses possibles :
- Python
- Ruby
- Java
- JavaScript
Explication : Python was created by Guido van Rossum and released in 1991.
Question 2 : Which company developed the GPT-4 AI model?
Réponses possibles :
- OpenAI
- Meta
- Microsoft
Explication : OpenAI, founded in 2015 by Sam Altman, Elon Musk, and others, launched GPT-4 in March 2023. This multimodal model can analyze text and images. OpenAI also created ChatGPT (November 2022), which reached 100 million users in just 2 months, an absolute record.
Question 3 : What does the acronym SQL stand for?
Réponses possibles :
- Standard Query Logic
- System Query Language
- Simple Query Language
- Structured Query Language
Explication : SQL signifie Structured Query Language, utilisé pour gérer les bases de données.
Question 4 : Which protocol secures web connections with a padlock?
Réponses possibles :
- FTP
- HTTPS
- SSH
- HTTP
Explication : HTTPS (HyperText Transfer Protocol Secure) encrypts web communications.
Question 5 : Which technology enables the creation of decentralized applications?
Réponses possibles :
- IoT
- Blockchain
- Cloud Computing
- Machine Learning
Explication : Blockchain technology enables the creation of secure decentralized applications (dApps).
Question 6 : Which Google algorithm, launched in 2015, uses deep learning to improve search results?
Réponses possibles :
- RankBrain
- TensorFlow
- PageRank
- DeepMind
Explication : RankBrain is an artificial intelligence system used by Google since 2015 to process search queries and improve result relevance.
Question 7 : Which company developed the Watson supercomputer that beat human champions at Jeopardy! in 2011?
Réponses possibles :
- Microsoft
- Apple
- IBM
Explication : IBM developed Watson, an AI system that won the TV show Jeopardy! in 2011 against champions Ken Jennings and Brad Rutter.
Question 8 : What term describes machine learning where the algorithm learns from labeled data?
Réponses possibles :
- Unsupervised learning
- Deep learning
- Supervised learning
- Reinforcement learning
Explication : Supervised learning uses labeled data (with correct answers) to train models, unlike unsupervised learning.
Question 9 : Which neural network architecture, invented in 2017, is the foundation of models like GPT and BERT?
Réponses possibles :
- CNN (Convolutional Neural Network)
- GAN (Generative Adversarial Network)
- Transformer
- RNN (Recurrent Neural Network)
Explication : The Transformer architecture, introduced in the paper "Attention is All You Need" in 2017, revolutionized natural language processing and is the basis of GPT, BERT, and Claude.
Question 10 : Which test created in 1950 evaluates whether a machine can imitate human thought?
Réponses possibles :
- Test de von Neumann
- Test de Lovelace
- Test de Turing
- Test de Shannon
Explication : The Turing test, proposed in 1950 in the paper "Computing Machinery and Intelligence", evaluates whether a machine can exhibit intelligent behavior indistinguishable from that of a human.
Question 11 : Which AI technique generates images from text descriptions like "an astronaut cat"?
Réponses possibles :
- OCR
- AI image generation
- Facial recognition
- Compression d'images
Explication : AI image generation (like DALL-E, Midjourney, Stable Diffusion) uses diffusion models to create images from text prompts.
Question 12 : What concept describes the risk that a superintelligent AI might pursue goals contrary to human interests?
Réponses possibles :
- Alignment problem
- Technological singularity
- Echo chamber effect
- Biais algorithmique
Explication : The AI alignment problem concerns the difficulty of ensuring that an advanced AI pursues beneficial goals aligned with human values.
Question 13 : Which Elon Musk company develops brain implants to connect the brain to a computer?
Réponses possibles :
- The Boring Company
- Tesla
- SpaceX
- Neuralink
Explication : Neuralink, founded by Elon Musk in 2016, develops brain implants to create direct interfaces between the human brain and computers.
Question 14 : What type of neural network is particularly effective for image recognition?
Réponses possibles :
- Réseau de Hopfield
- Multilayer perceptron
- Convolutional Neural Network (CNN)
- Recurrent Neural Network (RNN)
Explication : Convolutional neural networks (CNNs) are specially designed to process grid-like data such as images, using convolution layers that detect visual features.
Question 15 : What term describes an AI model's ability to explain its decisions in an understandable way?
Réponses possibles :
- IA symbolique
- IA faible
- IA générative
- IA explicable
Explication : Explainable AI (XAI) aims to make AI model decisions transparent and understandable to humans, unlike "black boxes".
Question 16 : Which open-source Google framework has become a standard for machine learning model development?
Réponses possibles :
- TensorFlow
- React
- Angular
- Flutter
Explication : TensorFlow, developed by Google Brain and made open-source in 2015, is one of the most popular frameworks for creating and training machine learning models.
Question 17 : What phenomenon occurs when an AI model memorizes training data instead of generalizing?
Réponses possibles :
- Hallucination
- Sous-apprentissage
- Dérive des données
- Overfitting
Explication : Overfitting occurs when a model adapts too closely to training data and loses its ability to generalize to new data.
Question 18 : Which technology allows autonomous cars to detect their environment in 3D using lasers?
Réponses possibles :
- Sonar
- GPS
- LiDAR
- Radar
Explication : LiDAR (Light Detection and Ranging) uses laser beams to create a detailed 3D map of the environment, essential for autonomous vehicle navigation.
Question 19 : What term describes synthetic content (videos, audio) created by AI to deceive viewers?
Réponses possibles :
- Morphing
- Deepfake
- Photomontage
- Fake news
Explication : Deepfakes use AI to create hyperrealistic synthetic content, such as videos where one person's face is replaced with another's.
Question 20 : Which quantum computing concept could threaten the security of current cryptography?
Réponses possibles :
- Artificial general intelligence
- Suprématie quantique
- Réalité virtuelle
- Internet des objets
Explication : Quantum supremacy and quantum algorithms like Shor's could break current RSA cryptography systems, hence the development of post-quantum cryptography.
