Engineers from China have unveiled a device designed to translate the sounds of domestic animals into comprehensible human speech. The smart gadget is being developed by the startup Meng Xiaoyi, based in Hangzhou.
The creators claim that the accuracy in recognizing dog barks and cat meows reaches 95 percent. This novel device is in the form of a special collar and has already garnered significant online attention, with users placing thousands of pre-orders at a price of $118 per unit. The apparatus was publicly showcased at the It Expo technology exhibition in May, and its manufacturer is Hangzhou ZhiChong Interactive Technology.
According to the developers, the device captures sound waves and registers even the slightest shifts in a pet’s vocal tonality. This audio information is then transmitted to a cloud-based storage system, where the multimodal neural network Qwen compares the sounds against a database of pre-programmed emotional patterns for cats and dogs.
However, external experts point out that it is impossible to fully train a neural network to understand animal language within a few months, suggesting that the project is still in its active development phase.
Previously, a team of scientists from Japan’s Iwate University conducted research to determine the reasons why cats might refuse their usual food, even when feeling very hungry.
About the Author
