Hi readers!
Hope you enjoyed the first blog post of the new series. Here is the second blog in which we will talk about details of Artificial intelligence (AI) powered tools: the most common among them is ChatGPT.
Chat GPT (Chat Generative Pre-Trained Transformer) is developed by OpenAI and launched on November 30th, 2022: an application that serves as a virtual assistant capable of engaging in human-like conversations with users.
Now, let us first discuss various terms used in Chat GPT.
Open AI (Not to be confused with open AL (Open Adio Library) is an American AI research organization founded in December 2015, for research on artificial intelligence (AI) with the objective to develop “safe and beneficial” artificial general intelligence, (AGI) defined as “highly autonomous system” that outperform humans at most economically valuable work.
GPT stands for Generative Pre-trained Transformer. It is a family of large language models created by OpenAI that uses deep learning method to generate human-like conversational text. GPT is a generative AI technology that has been previously trained to transform its input into a different type of output. Artificial neural networks are used in natural language processing tasks and are based on the transformer architecture pre-trained on large data sets of unlabeled text.
Deep learning is a subset of machine learning that uses multi-layered neural networks called deep neural networks (DNN) to simulate the complex decision-making power like of human brain.
A typical DNN) is a three or more-layered network (Generally, DNNs have many more layers). They are trained on large amounts of data to identify within it: phenomena, recognize patterns and relationships, evaluate possibilities, and make predictions and decisions. A single-layer neural network can make approximate predictions and decisions, but additional layers in DNN help refine and optimize those outcomes for greater accuracy.
Some form of deep learning methods powers most of the AI in our lives today.
Generative artificial intelligence (GAI) is capable of generating text, images, videos, and other data using generative models often in response to a prompts.
A prompt is a cue or a symbol that indicates the system’s readiness to perform the next command or the user’s current location. A prompt can also be a reminder or an incentive for action. It can be a text or a symbol on a computer screen, or a word or a gesture given to a performer.
Chat GPT was optimized for dialogue by using Reinforcement Learning with Human Feedback (RLHF): a method that uses human demonstration and preference to guide them towards desired behavior.
RLHF has two main components: Reinforcement Learning (RL) and Human Feedback (HF)and is about what exactly is being trained?
Historically, when we talk about LLM training, we just mean updating only the parameters of the language model.
For details, please visit:
https://www.assemblyai.com/blog/the-full-story-of-large-language-mod
Using the power of Natural Language Processing (NLP) and AI technology, ChatGPT goes beyond scripted responses to provide personalized interactions.
For example, ChatGPT can understand the context of conversations allowing it to generate truly relevant and accurate responses to the user’s queries.
Scripted responses are predetermined responses or phrases that are used in certain situations or interactions. They can be written for speeches, broadcasts, customer service, or other purposes. Scripted responses can limit the customer’s frustration, keep the interaction productive, save time, and lower labor costs. However, they can also be seen as impersonal, rigid, or unresponsive to the specific needs or questions of the audience or customer.
When and where Chat GPT goes beyond scripted responses, it indicates that it is preparing itself to listen to the customer’s questions and give appropriate replies.
By borrowing NLP, ChatGPT can adapt its responses based on the nuance (tones and tanner) of the language thereby creating a seamless and natural conversational flow. This ability to comprehend and respond effectively makes ChatGPT a valuable tool for information exchange and communication in various contexts.
Nuance means a very slight difference or variation in meaning, expression, sound, color, or feeling. Nuance is often used to describe the fine details or shades of meaning that are not easily noticeable or obvious.
I know dear readers; it is not easy to understand it unless you work on Chat GPT. Think about that because seeing is believing.
See you next week. Take care, bye.
