MACHINE LEARNING PATENTING
Patenting Machine Learning, ML, is turning out to be the hot topic of the day. While the users and the world celebrates the success of the benefits of ML, the undercurrent of commercial interests are moving in the direction of patenting, to keep their turf safe of any unwarranted commercial overtures from the known competitors and also the unknown Davids. Successful business models would reign supreme at the end of the day and patenting would be the bedrock of it.
The number of algorithms & techniques in ML are increasing by the day with researchers and engineers toiling hard towards its implementation & the getting the benefits thereof. The patent process is cumbersome the world over, only the degrees vary. There are legal & ethical issues as well. Though large number of technology companies are filing & getting patents, the issues related to it cannot be overlooked. Arguments are on both sides. Patents are intangible assets with powers of approval vested in the country concerned. US is supposed to be less rigorous than the EU, India too having a stringent patent filing process.
Google already has patents for Dropout, linear progression and word embedding, besides many others. Patient is at the core of making products user friendly simultaneously with commercial gain. The issue is whether it is a software, technology or a product, it has to prove its worth against existing line of products / services. Undeniably, lots of products are made open source by these and other companies as well. ”The patent ensures that computer-related interventions aren’t mere schemes, formulas or business methods.”
On the flip side, it is argued that ML discoveries do not focus on problem solving at a large scale, making this concept of patenting obsolete. It is also said that such solutions kill innovation by either hindering or obviating from the crux of the problem itself. Another fear is of broad patents being issued in AI space. Dropout is a case in point.
JUDICIOUS USE OF PATENTING ONLY, CAN SUSTAIN INNOVATION, WHILE RETAINING COMMERCIAL VALUE, WHEREVER WARRANTED.