Smaller AIs vs larger AIs.
Photo credit: theFreesheet/Google ImageFX

The industry’s reliance on building ever-larger models is becoming unsustainable due to “enormous computing resources” and energy costs, creating an urgent need for leaner artificial intelligence systems.

Researchers from Shanghai Jiao Tong University have outlined a comprehensive roadmap for “efficient multimodal large language models” that challenges the dominance of centralised cloud infrastructure. The review, published in Visual Intelligence, argues that the future of intelligence depends on reducing computational barriers rather than sheer scale.

“Efficiency determines who can build, deploy, and benefit from multimodal AI,” said Prof. Lizhuang Ma, the team leader of the study.

Critical flaw

The study identifies a critical flaw in current multimodal systems: visual inputs generate long token sequences that dramatically increase complexity. A single image can produce thousands of tokens, making standard models too heavy for practical deployment.

To solve this, researchers propose “vision token compression” to remove redundant data before it reaches the language model. They also advocate for compact language backbones with just one billion to three billion parameters, coupled with lightweight vision encoders.

Beyond simple compression, the review emphasises emerging architectures such as “mixture-of-experts”. These systems selectively activate specific model components to increase capacity without proportionally increasing computation costs.

This shift toward efficiency is expected to “democratise access” to advanced AI capabilities by allowing powerful models to run on mobile devices and edge platforms. The authors suggest this transition will enable real-time applications in healthcare and remote sensing while addressing growing concerns about energy consumption and data privacy.

Leave a Reply

Your email address will not be published. Required fields are marked *

You May Also Like

Massive AI study uncovers the secret GLP-1 side effects hidden on Reddit

Millions of patients are flocking to GLP-1 weight loss injections, but artificial…

Alarming new US survey shows half of patients rely on AI for medical choices

Across the United States, a dangerous new trend is emerging. Millions of…

One in four Americans now consult AI chatbots for medical advice

Millions of desperate patients are quietly abandoning the waiting room for a…

Global gambling firms rush to adopt AI despite severe lack of safety controls

The global gambling industry is racing to integrate artificial intelligence into its…

Why digital tears and online outrage fail to win modern political arguments

Scrolling through your social media feed today often feels like navigating a…

Tracking how war and energy policies dimmed night lights of Europe

While human civilisation is glowing brighter than ever before, the lights across…