High accuracy AI for malware classification

On Tuesday the paper Computer activity learning from system call time series that Curt and I wrote was posted to the Arxiv. It explains how we used machine learning to create a minute-by-minute description of what is happening on a computer.…

Keep reading

General malware detectors are impossible

It is impossible to write a general purpose malware detector. Not hard, not difficult, impossible. The argument for the impossibility relies on building an odd program. We may not write such a program in practice, but it does arise as a combination of things we do write — things like Perl-like regular expressions and input parsers — and carefully crafted inputs.…

Keep reading

More on the curse

The n-cube playground As a playground to understand the curse of dimensionality we spread 20,000 points throughout a 10-dimensional cube of side 2. Each coordinate of a point is independently chosen from a uniform random distribution ranging from -1 to 1.…

Keep reading

The two curses of dimensionality

The curse of dimensionality made its print appearance in Richard Bellman’s 1957 book Dynamic programming. It was an outcry over the impossibilities of dealing with functions of many variables when a computer with a million bytes of memory seemed beyond imagination.…

Keep reading

Given a wall, who wins?

Walls guard. Walls constrain. They defy us to break them. And so it has been since the dawn of agriculture. We see defensive walls protecting ancient cities from China through sub-Saharan Africa. Their widespread prevalence is a testment of their desirability.…

Keep reading