This browser is not actively supported anymore. For the best passle experience, we strongly recommend you upgrade your browser.
| 1 minute read

TinyML: The future for AI?

In the linked article, the author discusses a little known sub-branch of machine learning referred to as TinyML. TinyML is concerned with lightweight machine learning algorithms capable of running on a device, rather than on a server, exhibiting very low power consumption and memory overheads.

In many instances, due to the ever expanding model sizes, the use of machine learning models requires remote inference. In other words, the computation is performed in the cloud. In this case, the data provided as input to the model must be transmitted to the cloud for processing. Following the rise and widespread adoption of ChatGPT, questions arose regarding the confidentiality of such services, particularly in the context of commercially sensitive activity. Furthermore, services such as ChatGPT are a huge infrastructural burden to run. It has been reported that OpenAI spends approximately $700,000 a day to run ChatGPT, a large proportion likely spent on energy consumption.

In contrast, TinyML algorithms exhibit lower power consumption and offer enhanced integration of existing ML models on memory-restricted devices. This is achieved through a number of different techniques such as model optimization (e.g. parameter quantization, pruning, and knowledge distillation) and hardware acceleration. Furthermore, TinyML offers confidentiality baked in because the models are run locally and do not require data to be transmitted to third party providers.

In the world of patents and intellectual property in general, confidentiality is of the utmost importance. A breach of confidence or other public disclosure can prejudice an applicant’s right to gain protection for their idea. Furthermore, in industry, using machine learning services such as ChatGPT to process sensitive data may be in contravention of certain laws, treaties, or agreements. Therefore, the adoption of TinyML may be used to mitigate these risks.

Given TinyML’s reported benefits, it’s completely unsurprising that ABI Research, a global tech market advisory firm, projects that the TinyML market will grow from 15.2 million shipments in 2020 to 2.5 billion in 2030. Personally, I would echo the sentiment given in the linked article that TinyML has the power to positively transform lives, challenging initial fears about the risks associated with the rise of artificial intelligence.

For more information regarding patents and AI, please read our annual report on trends in AI at the European Patent Office. For specific information regarding the patentability of AI inventions, please feel free to get in touch as we'd be happy to help.

According to ABI Research, Tiny Machine Learning devices, requiring significantly less power than traditional machine learning models, are expected to grow from 15.2 million shipments in 2020 to 2.5 billion in 2030.


artificial intelligence, data & connectivity, digital transformation, internet of things, yes