Microsoft has created an artificial language that allows AI models to talk to each other faster and more efficiently than in English, with the hope that groups of models will be able to team up without having to resort to clumsy and sprawling human words.
Many researchers believe that using several artificial intelligence models, each with different specialisms and abilities, to solve problems collectively holds promise for tackling thorny problems that individual ones can鈥檛 solve. Although large language models like ChatGPT have been shown to be capable of communicating at high speed, even reaching consensus in groups of up to 1000, designing these models to talk to each other in a language like English 鈥� as humans would talk to them or each other 鈥� creates a significant bottleneck.
Now, at Microsoft and his colleagues听have created a language called Droidspeak to streamline that communication. Named after the beeping dialect 鈥渟poken鈥� by R2-D2 in the Star Wars听films, the language is essentially a mathematical shorthand for evoking certain words, concepts or instructions.听
Advertisement
Most current AI models take English prompts and split them into a series of tokens, which can be anything from individual characters to words or even whole phrases.听This string of tokens is then turned into a mathematical description of the complex path taken by each token as it is processed by the AI鈥檚 neural network. These paths embed conceptual and factual information about language.
For instance, the individual paths for the words 鈥渒ing鈥� and 鈥渜ueen鈥� will be more similar to each other than 鈥渒ing鈥� and 鈥渂all鈥� are. The paths can also be manipulated mathematically, such as by taking 鈥渒ing鈥�, subtracting 鈥渕an鈥�, adding 鈥渨oman鈥� and coming up with a path for 鈥渜ueen鈥�.
鈥淚t鈥檚 not working in words, it鈥檚 not working in tokens, it鈥檚 working in high-dimensional space,鈥� says at the University of Maryland, Baltimore County, who wasn鈥檛 involved in the work. 鈥淭hese are very complicated curves.鈥�
Droidspeak allows the AI models to remain working in this high-dimensional space, rather than converting into text and back, which the researchers say is 2.78 times faster and has negligible accuracy loss. But there is a catch: currently, it only works with multiple copies of the same AI model. The researchers say they are hoping to investigate how to apply the concept to models that differ from each other.
Feldman says the approach will speed up communication between AI models and could also make collections of models more powerful and able to tackle larger and more complex problems.
He points out that similar ideas have been used in the past to attack AI and get it to do things that the makers want to prevent. For instance, if a model checks inputs for dangerous keywords like 鈥渂omb recipe鈥�, you can鈥檛 get it to tell you how to make a bomb. But by working out the path for a sequence of words that conveys the same meaning, you may be able to input that to a model and sidestep the safety mechanisms.
arXiv
