so I'm the author of one of the first usable unum implementations. We're actually moving to something called sigmoid numbers which are even better. The lecture doesn't cover "valid mode" but that's more like the "unum". At the end of the video, there's a demonstration where I show some very interesting results concerning machine learning.
Also, I am working on some verilog implementations, so we may have this in hardware rather soon.
A few things about sigmoid numbers: 1) they have fixed sizes, so you don't have to worry about keeping track of things (which was a pain). 2) they have a property I call "isomorphic structure" - which means that zero-padding does not affect the value, and truncation gives you the nearest value at a lower bit resolution. 3) they really do perform better than floating points in several applications. There's a higher bit-per-bit entropy for 'most things you want to do', so the information content is richer. The distribution of the numbers is "more human" - that is to say at a certain point you stop caring about the details around the number, and more care that it's 'really big' or 'really small'; and the distribution reflects that.
> very interesting results concerning machine learning
This is the big chance for alternative number formats. a whole lot of people are working on neural net ASICs these days and IEEE 754 is not a requirement. A new number format with a low power/low area hardware implementation could easily find adoption in this new area.
no, not yet. john literally invented them either late december or early january. If you have any questions, feel free to ask me (contact info in profile). I'm also happy to send you a draft "article" that John is writing.
https://www.youtube.com/watch?v=aP0Y1uAA-2Y
You can try out sigmoid numbers in julia here:
https://github.com/interplanetary-robot/SigmoidNumbers
Also, I am working on some verilog implementations, so we may have this in hardware rather soon.
A few things about sigmoid numbers: 1) they have fixed sizes, so you don't have to worry about keeping track of things (which was a pain). 2) they have a property I call "isomorphic structure" - which means that zero-padding does not affect the value, and truncation gives you the nearest value at a lower bit resolution. 3) they really do perform better than floating points in several applications. There's a higher bit-per-bit entropy for 'most things you want to do', so the information content is richer. The distribution of the numbers is "more human" - that is to say at a certain point you stop caring about the details around the number, and more care that it's 'really big' or 'really small'; and the distribution reflects that.