samedi 11 mars 2023

GOUVERNANCE - AI: Parlez-vous ChatGPT ? - Do you speak ChatGPT ?

GOUVERNANCE - AI: Parlez-vous ChatGPT ? - Do you speak ChatGPT ?

Parlez-vous ChatGPT ? - Do you speak ChatGPT ?

Parlez-vous ChatGPT ? - Do you speak ChatGPT ?



• 2015. OpenAI was founded by Sam Altman, Elon Musk, Greg Brockman, Peter Thiel, and others. OpenAI develops many different AI models other than GPT.

• 2017. Google published the paper Attention is All You Need, which introduced the transformer architecture [2]. The transformer is a neural network architecture that lays the foundation for many state-of-the-art (SOTA) large language models (LLM) like GPT.

• 2018. GPT is introduced in Improving Language Understanding by Generative Pre-training . It’s based on a modified transformer architecture and pre-trained on a large corpus.

• 2019. GPT-2 is introduced in Language Models are Unsupervised Multitask Learners, which can perform a range of tasks without explicit supervision when training.

• 2020. GPT-3 is introduced in Language Models are Few-Shot Learners, which can perform well with few examples in the prompt without fine-tuning.

• 2022. InstructGPT is introduced in Training language models to follow instructions with human feedback, which can better follow user instructions by fine-tuning with human feedback.

• 2022. ChatGPT, a sibling of InstructGPT, is introduced in ChatGPT: Optimizing Language Models for Dialogue. It can interact with humans in conversations, thanks to the fine-tuning with human examples and reinforcement learning from human feedback (RLHF).


Natural language processing (NLP) 

The transformer architecture is the foundation for GPT. 


The first step is supervised learning from human examples. 


The second step is training a reward model (RM) to rate the responses

Optimizing Language Models for Dialogue.


NLP Using Python     Natural Language Processing


Machine Learning Mastery



APA handbook of industrial and organizational psychology, Vol 1: Building and developing the organization. 


Definition of prompt for ChatGPT in English with APA and exemples and URL 

The term "Prompt" in English can have several meanings depending on the context. In the field of artificial intelligence, the term "Prompt" can be translated as "prompt" or "proposal". PromptBase | Prompt Marketplace: DALL·E, Midjourney, ChatGPT, Stable Diffusion & GPT-3.


As an AI language model, I don't have a personal definition for the term prompt. However, here is a definition with APA and examples:

Prompt is a noun that refers to a cue or stimulus that initiates or triggers a particular action, behavior, or response. The term is commonly used in education to describe a task or question given to students that is designed to encourage critical thinking, creativity, and reflection.

According to the American Psychological Association (APA), a prompt is defined as "a stimulus or cue that initiates or guides behavior, thought, or action" (APA Dictionary of Psychology, 2020).

Examples of prompts include:

In a writing class, the teacher might provide a prompt such as "Write a descriptive essay about your favorite place" to inspire students to write creatively and think deeply about their experiences.

In a language learning class, the teacher might give a prompt such as "Describe your daily routine in Spanish" to encourage students to practice their language skills and use new vocabulary.

In a behavioral therapy session, the therapist might use a prompt such as "When you feel anxious, try taking three deep breaths and saying a positive affirmation to yourself" to help the patient develop coping skills and manage their emotions.

Here is the reference for the APA Dictionary of Psychology:

American Psychological Association. (2020). APA Dictionary of Psychology (2nd ed.). https://doi.org/10.1037/12169-000



Models - OpenAI API


The OpenAI API can be applied to virtually any task that involves understanding or generating natural language or code. We offer a spectrum of models with different levels of power suitable for different tasks, as well as the ability to fine-tune your own custom models. These models can be used for everything from content generation to semantic search and classification.

INFOS and DEMOS + PPT





--- 
 Pierre Erol GIRAUDY 


Aucun commentaire:

Enregistrer un commentaire

Des logiciels de reconnaissance de texte pour la généalogie avec l'IA

Voici un outil de reconnaissance de texte. Cependant, pour la généalogie, le logiciel Transkribus semble être le plus approprié. Transkribu...