At Google I/O, Google kicks off the day with a spotlight on AI — and its new generative artificial intelligence (AI) language model, PaLM 2. The state-of-the-art language model is a powerful addition to the field of AI, with significantly improved capabilities over previous versions.
PaLM 2 has been trained on Google’s Tensor Processing Unit, and excels at tasks such as mathematics, coding, reasoning, multilingual translation, and natural language generation. It can also understand nuanced phrases in different languages, while delivering stronger logic, common sense reasoning capabilities, and more efficient inference than its predecessors. As well, PaLM 2 can be fine-tuned to accomplish particular tasks, allowing it to power and work within over 25 Google products and services, such as Google’s product suite, Workspace, Cloud, Security, and Vertex AI.
To celebrate PaLM 2’s launch, Google has organized a gathering for top executives in San Francisco for July 11-12. Here, leaders will be able to gain insight on how to leverage and maximize investments in AI for success.
Zoubin Ghahramani, VP of Google DeepMind, has spearheaded the development of PaLM 2. In a roundtable press conference, Ghahramani expressed his passion for building products and services that make the world’s information more accessible, which has been given deeper meaning through the advent of AI.
Google is at the forefront of the market with the launch of PaLM 2, as companies like Microsoft and OpenAI are seeing the successes of ChatGPT. By training the model and refining its capabilities, Google is aiming to set the tone for the industry, driving the conversation forward and providing greater advancements in AI innovation.