Google ups ante in GenAI with Gemini enhancements | Computer ...
Google has fired the latest salvo in the race for artificial intelligence (AI) supremacy with significant enhancements to its Gemini model, including a groundbreaking two-million-token context window for Gemini 1.5 Pro.
Gemini 1.5 Pro Enhancements
Gemini 1.5 Pro, Google’s multimodal generative AI (GenAI) model, can now analyse and classify video, audio, code, and text. This advancement enables applications like chatbots to effectively handle complex scenarios involving various content types. For instance, processing motor claims with related video and textual evidence.
Launched earlier this year with a one-million-token context window, the model now boasts double the capacity. This enhancement enables it to process significantly more information, including analysing 30,000 lines of code or uploading entire database tables and schemas for streamlined SQL analysis.
Next-Gen Context Window
The new enhancement, currently available through a waitlist for developers, goes beyond simply handling large volumes of data. Google aims to achieve an unlimited context window size in the future for smarter, more comprehensive interactions with information.
Context Caching and Efficiency
Google will also introduce context caching to Gemini 1.5 Pro next month. This feature will allow users to send large files and other parts of a prompt only once, making the expansive context window more useful and cost-effective.
Gemini 1.5 Pro's ability to handle larger context windows is attributed to Google’s Mixture-of-Experts (MoE) architecture. This architecture increases model capacity without a proportional increase in computation, eliminating the need to fine-tune foundation models extensively.
Incorporating RAG and Gemini 1.5 Flash
While Google's Mixture-of-Experts (MoE) architecture enhances the model's capacity, the importance of retrieval augmented generation (RAG) in refining output accuracy and relevance, specifically for coding applications, remains crucial.
For applications prioritising low latency and cost efficiency, Google has introduced Gemini 1.5 Flash. This model is optimised for narrower or high-frequency tasks where rapid response times are critical.
According to Demis Hassabis, CEO of Google DeepMind, Gemini 1.5 Flash "excels at summarisation, chat applications, image and video captioning, data extraction from long documents and tables, and more" due to its distillation process.
At the MIT Sloan CIO Symposium, IT leaders discussed the challenge of managing the "explosion" of GenAI companies while also highlighting the importance of cybersecurity regulations for businesses.
Google continues to drive innovation in the AI space with these latest Gemini enhancements, offering new possibilities for developers and businesses alike.
