The three big AI trends for 2024

2023 was a year in which generative Artificial Intelligence (AI) made great strides, moving from research labs to real life, with millions of people using it through popular tools such as ChatGPT and Microsoft Copilot. By 2024, AI is expected to become more accessible and integrated into technologies that improve everyday tasks and help solve difficult problems. 

Based on this, Microsoft has defined the following three major AI trends for 2024:

• Small Language Models (SLM)

Large language models (LLMs) like Copilot require many computing resources, making small language models (SLMs) a relevant alternative. Although still large, with lots of parameters, SLMs can be run offline on a cell phone. These parameters are variables that determine the model's behavior.

Sebastien Bubeck from Microsoft Research says that SLMs can make AI more accessible due to their size and cost and that ways are being developed to make them as powerful as LLMs.

Microsoft has developed the Phi and Orca SLMs, which perform comparably to or better than LLMs in certain areas, challenging the idea that scale is essential for performance. Unlike LLMs, which use large amounts of data from the Internet, SLMs use selected, high-quality training data, allowing them to reach new limits in terms of size and performance. In the coming months, improved models are expected to be launched to promote further research and innovation.
 

• Multimodal AI

Most LLMs only process text, but multimodal models can understand text, images, audio, and video, enriching and making more accurate technologies such as search tools and creativity applications.

Copilot uses a multimodal model to process images, natural language, and Bing search data, making it possible to identify a monument's historical significance in an uploaded photograph.

Multimodal AI also powers Microsoft Designer, which generates images from textual descriptions, and makes it possible to create personalized neural voices, useful in text readers and for people with vocal disabilities.

“Multimodality can create more human experiences, taking better advantage of senses such as sight, speech, and hearing,” says Jennifer Marsman, principal engineer in the Office of the Chief Technology Officer at Microsoft, Kevin Scott.
 

• AI in science

Experts predict major advances in AI tools to accelerate scientific discoveries, focused on solving global problems such as climate change, energy crises, and diseases.

To mitigate climate change and increase agricultural efficiency, Microsoft researchers use AI to create better weather forecasts, carbon estimators, and tools for sustainable agriculture. 

In life sciences, researchers collaborate on the largest image-based AI model to fight cancer and use AI to discover new drugs for infectious diseases and molecules for innovative treatments.

AI is transforming materials science, making creating new materials with specific properties easier. A recent discovery showed how AI and high-performance computing can accelerate the search for less toxic battery materials.

“Artificial intelligence is driving a revolution in scientific discovery,” says Chris Bishop, director of the AI4Science team at Microsoft Research. “This could be the most exciting and important application of AI.”.


Follow Hydra iT to keep up to date with all the latest technology! Contact us to transform your business!

^{Source: Microsoft (https://news.microsoft.com/three-big-ai-trends-to-watch-in-2024/?ocid=FY24_soc_omc_br_li_Trends)}