Immediately after the announcement of Gemini, Google's advanced language model, Bard it was promptly updated to incorporate the improvements of Gemini Pro, initially limited to the English version of the chatbot.
A few months later, the company announced the extension of its artificial intelligence model to the virtual assistant in additional countries
The big G has expanded the reach of Gemini Pro, adding more 40 languages following its initial introduction last December.
Among the languages involved there is also Italian and, furthermore, a name change is planned for the Google chatbot itself.
Pit stops
The Bard chatbot, similar to OpenAI's ChatGPT, but based on the LLM GPT-3.5 or GPT-4, previously powered by LaMDA, a unimodal model, has been linked to PaLM 2, multimodal but less powerful than Gemini Pro, since May 2023.
Gemini includes various models, including Gemini Ultrathe largest and most powerful designed for more complex tasks, Gemini Dwarf divided into Nano-1 and Nano-2 for mobile devices like Pixel 8 Pro, and Gemini Proversatile for a wide range of activities.
Gemini models are multimodalthey therefore accept input of text, images, audio, video and programming code.
With this update, target users can benefit from the advanced features introduced by Gemini in Bard, relating to understanding, reasoning, synthesis and writing code in programming tasks, as well as new features previously unavailable.
Recent independent analyzes have recognized Bard with Gemini Pro as a more advanced chatbot than OpenAI's GPT-4-based competitors, such as ChatGPT or Microsoft Copilot.
Additionally, Bard will now feature a“G” icon (from the Google logo) visible at the end of the answers.
This feature represents a double check function that allows users, by tapping it, to initiate a process of confirming the validity of answers through a web search, automatically highlighting affected parts of the answer and allowing users to view confirmatory or conflicting information.
Multifunctional chatbot
Regarding a potential name change, Google seems inclined to, according to Android app developer Dylan Roussel rename Bard as “Gemini”establishing a direct connection with the model that powers it.
The name change was mentioned in the Bard web app changelog.
Additionally, Google announced the rollout of new AI-based features for Bard.
Among these, the ability to generate images from textwith the possibility of changing the style intuitively.
The ability to generate images from text descriptions, currently only available for the English version of the English assistant, is made possible thanks to the Imagen 2 template.
A further update concerns a music generatorwhich offers users the ability to select the style of the song, its duration and the purpose for which it is intended.
Importantly, any content generated via these features will be marked with the SynthID “digital mark”, allowing for verification of origin via artificial intelligence.
#Google #enhances #Bard #Gemini #Pro #multimodal #model #Italy
